From matthew_pocock at yahoo.co.uk  Mon Mar  1 06:42:51 2004
From: matthew_pocock at yahoo.co.uk (Matthew Pocock)
Date: Mon Mar  1 06:49:30 2004
Subject: [Biojava-l] Bug in biojava GUI?
In-Reply-To: <200402290833.i1T8XF4q003989@mail.cbi.pku.edu.cn>
References: <200402290833.i1T8XF4q003989@mail.cbi.pku.edu.cn>
Message-ID: <404321BB.1080900@yahoo.co.uk>

Hi Wux,

Sounds like a bug to me. Could you send me some example code? I will try
to track it down. What version of Java and BioJava are you using?

Matthew

wux@mail.cbi.pku.edu.cn wrote:

>Dear all:
>
>         When I add a TitleBorder to SequencePanel, The bottom line in MutilLineRenderer is disappear. I change the border toLineBorder, all the lines are in the line frame. I am not sure it is due to biojava's problem or java's . 
>For example: 
>      ---- Title Border --------------------------
>      |    ----->     ----->                     |
>      |        ----->                            |
>      |    <----                                 |
>      |  acgtttttttttaaatttttttttttttttttttttttt |
>      --------------------------------------------
>         ----5------10------15-------------------- ( This line is disappear !)
>
>change a border to SequencePanel :
>      ----------------- --------------------------
>      |    ----->     ----->                     |
>      |        ----->                            |
>      |    <----                                 |
>      |  acgtttttttttaaatttttttttttttttttttttttt |
>      |  ----5------10------15-------------------| ( This line is correct in line frame !)
>      --------------------------------------------
>¡¡¡¡
>  Who else meets the same problem?
>¡¡¡¡¡¡¡¡¡¡¡¡              
> 				          Yours faithfully,
>¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡     wux
>¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡     wux@mail.cbi.pku.edu.cn
>¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡    2004-02-29
>*****************************************************
>WuXin  Ph.D student of CBI (Center of Bioinformatics)
>Peking University  100871 P.R.China
>Email: wux@mail.cbi.pku.edu.cn
>Tel: 010-62762409 (dorm)
>     010-62755206 (office)
>Address: Building 47#2026 Peking University
>*****************************************************
>
>
>_______________________________________________
>Biojava-l mailing list  -  Biojava-l@biojava.org
>http://biojava.org/mailman/listinfo/biojava-l
>
>  
>

From rohini_sulatycki at yahoo.com  Mon Mar  1 11:50:18 2004
From: rohini_sulatycki at yahoo.com (Sulatycki Rohini)
Date: Mon Mar  1 11:56:12 2004
Subject: [Biojava-l] Biojava development help needed? 
In-Reply-To: <OF7A02A414.94813A4C-ON48256E3C.001D8749-48256E3C.001D7240@EU.novartis.net>
Message-ID: <20040301165019.81904.qmail@web9808.mail.yahoo.com>

Hello,

I am a newbie to Biojava and would like to contribute
in any way to the project. I am fairly experienced
Java architect/developer and can contribute towards
development, architecture etc.

I would appreciate it if this group could let me know
if/how I can help. 

Thanks
Rohini Sulatycki

=====
Rohini Sulatycki
Technical Architect
rohini_sulatycki@yahoo.com

__________________________________
Do you Yahoo!?
Get better spam protection with Yahoo! Mail.
http://antispam.yahoo.com/tools
From douglas.hoen at mail.mcgill.ca  Thu Mar  4 13:33:18 2004
From: douglas.hoen at mail.mcgill.ca (Douglas Hoen)
Date: Thu Mar  4 13:39:28 2004
Subject: [Biojava-l] NPE from FlatModel.<init> with SimpleModelInState
Message-ID: <1078425198.2996.215.camel@elegans>

Hi,

I want to put a two-headed markov model in a SimpleModelInState.
But when I attempt to call DPFactory.DEFAULT.createDP, I get a
NullPointerException from inside the org.biojava.bio.dp.FlatModel
constructor. 

I get the same problem from the biojava-1.30-jdk13.jar,
biojava-1.30-jdk14.jar, and biojava-1.3.1.jar distributions.

Also, I turned up a similar thread in Dec. 2002 biojava-l archive
entitled "SimpleModelInState Problem?". Matthew Pocock responded to the
query saying he was about to work on the issue, but there is no
followup.

My questions: Is this a real bug? If so, should I expect this API
(ModelInState for two-headed models) to be supported in the near future,
or should I simply build a flat model?

The stack trace, possible problem biojava code, and snippets of my test
code are below. 

Thanks very much for any help.
Doug


Here is the stack trace:
------------------------

Exception in thread "main" java.lang.NullPointerException
        at org.biojava.bio.dp.FlatModel.<init>(FlatModel.java:251)
        at org.biojava.bio.dp.DP.flatView(DP.java:168)
        at
org.biojava.bio.dp.DPFactory$DefaultFactory.createDP(DPFactory.java:52)
        at ltr.ModelInStateExample.main(ModelInStateExample.java:124)
rethrown as org.biojava.utils.NestedError: Can't align.
        at ltr.ModelInStateExample.main(ModelInStateExample.java:133)


The NPE seems to be coming from the following FlatModel code:
-------------------------------------------------------------

//
// FIXME -- [SOMEONE (ed.)] broked this...
//

TranslatedDistribution dist = null;
// TranslatedDistribution dist TranslatedDistribution.getDistribution(
//  delegate.transitionsFrom(s),
//  sModel.getWeights(sOrig)
// );
SimpleReversibleTranslationTable table =
	(SimpleReversibleTranslationTable) dist.getTable();


Here is the model creation code:
--------------------------------

...
FiniteAlphabet singleAlphabet = DNATools.getDNA();
FiniteAlphabet doubleAlphabet = 
    ( FiniteAlphabet ) AlphabetManager.getCrossProductAlphabet( 
        Collections.nCopies( 2, singleAlphabet ) );

State matchState = new SimpleEmissionState(
        "match",
        Annotation.EMPTY_ANNOTATION,
        new int[] { 1, 1 },
        new UniformDistribution( doubleAlphabet )
        );
State insertState = new SimpleEmissionState(
        "insert",
        Annotation.EMPTY_ANNOTATION,
        new int[] { 0, 1 },
        new PairDistribution( 
            new GapDistribution( singleAlphabet ),
            new UniformDistribution( singleAlphabet )
        ) );
State deleteState = new SimpleEmissionState(
        "delete",
        Annotation.EMPTY_ANNOTATION,
        new int[] { 1, 0 },
        new PairDistribution(                     
            new UniformDistribution( singleAlphabet ),
            new GapDistribution( singleAlphabet )
        ) );

SimpleMarkovModel innerModel = 
    new SimpleMarkovModel( 2, doubleAlphabet, "inner");
innerModel.addState( matchState );
innerModel.addState( insertState );
innerModel.addState( deleteState );

innerModel.createTransition( innerModel.magicalState(), matchState );
innerModel.createTransition( matchState, insertState );
innerModel.createTransition( matchState, deleteState );
innerModel.createTransition( matchState, matchState );
innerModel.createTransition( matchState, innerModel.magicalState() );
innerModel.createTransition( insertState, matchState );
innerModel.createTransition( insertState, insertState );
innerModel.createTransition( deleteState, matchState );
innerModel.createTransition( deleteState, deleteState );

double d = 0.1;
double e = 0.1;
double t = 0.1;

innerModel.getWeights( innerModel.magicalState() )
            .setWeight( matchState,  1.0 );

Distribution fromMatch = innerModel.getWeights( matchState );
fromMatch.setWeight( insertState, d );
fromMatch.setWeight( deleteState, d );
fromMatch.setWeight( matchState, 1-2*d-t );
fromMatch.setWeight( innerModel.magicalState(), t );

Distribution fromInsert = innerModel.getWeights( insertState );
fromInsert.setWeight( matchState, 1-e );
fromInsert.setWeight( insertState, e );

Distribution fromDelete = innerModel.getWeights( deleteState );
fromDelete.setWeight( matchState, 1-e );
fromDelete.setWeight( deleteState, e );

State innerModelInState = 
    new SimpleModelInState( innerModel, "modelInState" );

SimpleMarkovModel outerModel = 
    new SimpleMarkovModel( 2, doubleAlphabet, "outer" );

outerModel.addState( innerModelInState );

outerModel.createTransition( 
    outerModel.magicalState(), innerModelInState );
outerModel.createTransition( 
    innerModelInState, outerModel.magicalState() );

outerModel.getWeights( outerModel.magicalState() )
            .setWeight( innerModelInState, 1.0 );
outerModel.getWeights( innerModelInState )
            .setWeight( outerModel.magicalState(), 1.0 );

...

DPFactory.DEFAULT.createDP( outerModel );  <==================== NPE


From Gera.Jellema at wur.nl  Thu Mar  4 08:30:12 2004
From: Gera.Jellema at wur.nl (Jellema, Gera)
Date: Thu Mar  4 19:36:29 2004
Subject: [Biojava-l] Reading frames and amino acids
Message-ID: <EF1BA7172DC75F4182B8E8A5B08EF6400E5ED5@salte0008.wurnet.nl>

Hi,
I'm new to biojava so I don't know if this has already been asked. I have a genome sequence and I want to have it in the 6 reading frames, and then per reading frame translated into amino acids so I can look for proteins. I don't know how I have to do it.
Thanks,
Gera
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/ms-tnef
Size: 2515 bytes
Desc: not available
Url : http://portal.open-bio.org/pipermail/biojava-l/attachments/20040304/553abd15/attachment.bin
From mark.schreiber at group.novartis.com  Thu Mar  4 23:16:09 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Thu Mar  4 23:19:26 2004
Subject: [Biojava-l] Reading frames and amino acids
Message-ID: <OF76CB5572.1634C3EF-ON48256E4E.00173462-48256E4E.00172152@EU.novartis.net>

Hi Gera -

The code below seems to work for me. Please note that I have not tested it 
thoroughly so you might want to eyeball a few results to make sure they 
are sensible.

You could probably improve it by making some of the 'in line' code into 
methods etc.

- Mark

import java.io.*;

import org.biojava.bio.*;
import org.biojava.bio.seq.*;
import org.biojava.bio.seq.io.*;
import org.biojava.bio.symbol.*;

/**
 * <p>Program to six-frame translate a nucleotide sequence</p>
 */

public class Hex {
  /**
   * Call this to get usage info, program terminates after call.
   */
  public static void help() {
    System.out.println(
        "usage: java utils.Hex <file> <format eg fasta> <dna|rna>");
    System.exit( -1);
  }
 
  public static void main(String[] args) throws Exception{
    if (args.length != 3) {
      help();
    }
 
    BufferedReader br = null;
    String format = args[1];
    String alpha = args[2];
 
    try {
      br = new BufferedReader(new FileReader(args[0]));
 
      SequenceIterator seqi = 
          (SequenceIterator)SeqIOTools.fileToBiojava(format, alpha, br);
 
      //for each sequence
      while(seqi.hasNext()){
        Sequence seq = seqi.nextSequence();
 
        //for each frame
        for (int i = 0; i < 3; i++) {
          SymbolList prot;
          Sequence trans;
 
          //take the reading frame
          SymbolList syms = seq.subList(
                i+1,
                seq.length() - (seq.length() - i)%3);


          //if it is DNA transcribe it to RNA 
          if(syms.getAlphabet() == DNATools.getDNA()){
            syms = RNATools.transcribe(syms);
          }
 
          //output forward translation to STDOUT
          prot = RNATools.translate(syms);
          trans = SequenceTools.createSequence(prot, "", 
                                               seq.getName()+
                                               "TranslationFrame: +"+i,
 Annotation.EMPTY_ANNOTATION);
          SeqIOTools.writeFasta(System.out, trans);
 
          //output reverse frame translation to STDOUT
          syms = RNATools.reverseComplement(syms);
          prot = RNATools.translate(syms);
          trans = SequenceTools.createSequence(prot, "",
                                               seq.getName() +
                                               "TranslationFrame: -" + i,
 Annotation.EMPTY_ANNOTATION);
          SeqIOTools.writeFasta(System.out, trans);
        }
      }
    }
    finally {
      if(br != null){
        br.close();
      }
    }
  }
}





"Jellema, Gera" <Gera.Jellema@wur.nl>
Sent by: biojava-l-bounces@portal.open-bio.org
03/04/2004 09:30 PM

 
        To:     <biojava-l@portal.open-bio.org>
        cc: 
        Subject:        [Biojava-l] Reading frames and amino acids


Hi,
I'm new to biojava so I don't know if this has already been asked. I have 
a genome sequence and I want to have it in the 6 reading frames, and then 
per reading frame translated into amino acids so I can look for proteins. 
I don't know how I have to do it.
Thanks,
Gera

_______________________________________________
Biojava-l mailing list  -  Biojava-l@biojava.org
http://biojava.org/mailman/listinfo/biojava-l

[ Attachment ''WINMAIL.DAT'' removed by Mark Schreiber ]


From dmb at mrc-dunn.cam.ac.uk  Fri Mar  5 04:29:53 2004
From: dmb at mrc-dunn.cam.ac.uk (Dan Bolser)
Date: Fri Mar  5 04:33:29 2004
Subject: [Biojava-l] Reading frames and amino acids
In-Reply-To: <OF76CB5572.1634C3EF-ON48256E4E.00173462-48256E4E.00172152@EU.novartis.net>
Message-ID: <Pine.LNX.4.21.0403050928200.18124-100000@mail.mrc-dunn.cam.ac.uk>



I have seen 'biojava in anger', but is their a 'biojava cookbook'?

It strikes me that lots of small scripts like this could be useful for
lots of people if archived properly.

Does the cookbook exist?

Cheers, Dan.

On Fri, 5 Mar 2004 mark.schreiber@group.novartis.com wrote:

> Hi Gera -
> 
> The code below seems to work for me. Please note that I have not tested it 
> thoroughly so you might want to eyeball a few results to make sure they 
> are sensible.
> 
> You could probably improve it by making some of the 'in line' code into 
> methods etc.
> 
> - Mark
> 
> import java.io.*;
> 
> import org.biojava.bio.*;
> import org.biojava.bio.seq.*;
> import org.biojava.bio.seq.io.*;
> import org.biojava.bio.symbol.*;
> 
> /**
>  * <p>Program to six-frame translate a nucleotide sequence</p>
>  */
> 
> public class Hex {
>   /**
>    * Call this to get usage info, program terminates after call.
>    */
>   public static void help() {
>     System.out.println(
>         "usage: java utils.Hex <file> <format eg fasta> <dna|rna>");
>     System.exit( -1);
>   }
>  
>   public static void main(String[] args) throws Exception{
>     if (args.length != 3) {
>       help();
>     }
>  
>     BufferedReader br = null;
>     String format = args[1];
>     String alpha = args[2];
>  
>     try {
>       br = new BufferedReader(new FileReader(args[0]));
>  
>       SequenceIterator seqi = 
>           (SequenceIterator)SeqIOTools.fileToBiojava(format, alpha, br);
>  
>       //for each sequence
>       while(seqi.hasNext()){
>         Sequence seq = seqi.nextSequence();
>  
>         //for each frame
>         for (int i = 0; i < 3; i++) {
>           SymbolList prot;
>           Sequence trans;
>  
>           //take the reading frame
>           SymbolList syms = seq.subList(
>                 i+1,
>                 seq.length() - (seq.length() - i)%3);
> 
> 
>           //if it is DNA transcribe it to RNA 
>           if(syms.getAlphabet() == DNATools.getDNA()){
>             syms = RNATools.transcribe(syms);
>           }
>  
>           //output forward translation to STDOUT
>           prot = RNATools.translate(syms);
>           trans = SequenceTools.createSequence(prot, "", 
>                                                seq.getName()+
>                                                "TranslationFrame: +"+i,
>  Annotation.EMPTY_ANNOTATION);
>           SeqIOTools.writeFasta(System.out, trans);
>  
>           //output reverse frame translation to STDOUT
>           syms = RNATools.reverseComplement(syms);
>           prot = RNATools.translate(syms);
>           trans = SequenceTools.createSequence(prot, "",
>                                                seq.getName() +
>                                                "TranslationFrame: -" + i,
>  Annotation.EMPTY_ANNOTATION);
>           SeqIOTools.writeFasta(System.out, trans);
>         }
>       }
>     }
>     finally {
>       if(br != null){
>         br.close();
>       }
>     }
>   }
> }
> 
> 
> 
> 
> 
> "Jellema, Gera" <Gera.Jellema@wur.nl>
> Sent by: biojava-l-bounces@portal.open-bio.org
> 03/04/2004 09:30 PM
> 
>  
>         To:     <biojava-l@portal.open-bio.org>
>         cc: 
>         Subject:        [Biojava-l] Reading frames and amino acids
> 
> 
> Hi,
> I'm new to biojava so I don't know if this has already been asked. I have 
> a genome sequence and I want to have it in the 6 reading frames, and then 
> per reading frame translated into amino acids so I can look for proteins. 
> I don't know how I have to do it.
> Thanks,
> Gera
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 
> [ Attachment ''WINMAIL.DAT'' removed by Mark Schreiber ]
> 
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 

From mark.schreiber at group.novartis.com  Fri Mar  5 04:35:14 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Fri Mar  5 04:38:56 2004
Subject: [Biojava-l] Reading frames and amino acids
Message-ID: <OF72E6EE9E.C4DC647A-ON48256E4E.00348FBC-48256E4E.003457E6@EU.novartis.net>

That is the intention of the 'biojava in anger' site. I just haven't had 
time to add lots of little scripts to it. I will gratefully take 
donations.

- Mark






Dan Bolser <dmb@mrc-dunn.cam.ac.uk>
03/05/2004 05:29 PM

 
        To:     Mark Schreiber/GP/Novartis@PH
        cc:     "Jellema, Gera" <Gera.Jellema@wur.nl>, biojava-l@portal.open-bio.org, 
biojava-l-bounces@portal.open-bio.org
        Subject:        Re: [Biojava-l] Reading frames and amino acids




I have seen 'biojava in anger', but is their a 'biojava cookbook'?

It strikes me that lots of small scripts like this could be useful for
lots of people if archived properly.

Does the cookbook exist?

Cheers, Dan.

On Fri, 5 Mar 2004 mark.schreiber@group.novartis.com wrote:

> Hi Gera -
> 
> The code below seems to work for me. Please note that I have not tested 
it 
> thoroughly so you might want to eyeball a few results to make sure they 
> are sensible.
> 
> You could probably improve it by making some of the 'in line' code into 
> methods etc.
> 
> - Mark
> 
> import java.io.*;
> 
> import org.biojava.bio.*;
> import org.biojava.bio.seq.*;
> import org.biojava.bio.seq.io.*;
> import org.biojava.bio.symbol.*;
> 
> /**
>  * <p>Program to six-frame translate a nucleotide sequence</p>
>  */
> 
> public class Hex {
>   /**
>    * Call this to get usage info, program terminates after call.
>    */
>   public static void help() {
>     System.out.println(
>         "usage: java utils.Hex <file> <format eg fasta> <dna|rna>");
>     System.exit( -1);
>   }
> 
>   public static void main(String[] args) throws Exception{
>     if (args.length != 3) {
>       help();
>     }
> 
>     BufferedReader br = null;
>     String format = args[1];
>     String alpha = args[2];
> 
>     try {
>       br = new BufferedReader(new FileReader(args[0]));
> 
>       SequenceIterator seqi = 
>           (SequenceIterator)SeqIOTools.fileToBiojava(format, alpha, br);
> 
>       //for each sequence
>       while(seqi.hasNext()){
>         Sequence seq = seqi.nextSequence();
> 
>         //for each frame
>         for (int i = 0; i < 3; i++) {
>           SymbolList prot;
>           Sequence trans;
> 
>           //take the reading frame
>           SymbolList syms = seq.subList(
>                 i+1,
>                 seq.length() - (seq.length() - i)%3);
> 
> 
>           //if it is DNA transcribe it to RNA 
>           if(syms.getAlphabet() == DNATools.getDNA()){
>             syms = RNATools.transcribe(syms);
>           }
> 
>           //output forward translation to STDOUT
>           prot = RNATools.translate(syms);
>           trans = SequenceTools.createSequence(prot, "", 
>                                                seq.getName()+
>                                                "TranslationFrame: +"+i,
>  Annotation.EMPTY_ANNOTATION);
>           SeqIOTools.writeFasta(System.out, trans);
> 
>           //output reverse frame translation to STDOUT
>           syms = RNATools.reverseComplement(syms);
>           prot = RNATools.translate(syms);
>           trans = SequenceTools.createSequence(prot, "",
>                                                seq.getName() +
>                                                "TranslationFrame: -" + 
i,
>  Annotation.EMPTY_ANNOTATION);
>           SeqIOTools.writeFasta(System.out, trans);
>         }
>       }
>     }
>     finally {
>       if(br != null){
>         br.close();
>       }
>     }
>   }
> }
> 
> 
> 
> 
> 
> "Jellema, Gera" <Gera.Jellema@wur.nl>
> Sent by: biojava-l-bounces@portal.open-bio.org
> 03/04/2004 09:30 PM
> 
> 
>         To:     <biojava-l@portal.open-bio.org>
>         cc: 
>         Subject:        [Biojava-l] Reading frames and amino acids
> 
> 
> Hi,
> I'm new to biojava so I don't know if this has already been asked. I 
have 
> a genome sequence and I want to have it in the 6 reading frames, and 
then 
> per reading frame translated into amino acids so I can look for 
proteins. 
> I don't know how I have to do it.
> Thanks,
> Gera
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 
> [ Attachment ''WINMAIL.DAT'' removed by Mark Schreiber ]
> 
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 




From off2w0rk at yahoo.com  Fri Mar  5 22:09:18 2004
From: off2w0rk at yahoo.com (facemann)
Date: Sat Mar  6 02:31:22 2004
Subject: [Biojava-l] Reading frames and amino acids
Message-ID: <20040306030918.83686.qmail@web41313.mail.yahoo.com>

Here is a small contribution.  I use it to find simple motifs. Feel free to edit or scrap.
 
/**
 *MotifLister.java
 *Andy Hammer
 *08 Aug 2003
 *Lists all instances of a motif in specified (dna\rna\protein) fasta file.
 *The motif can contain Ambiguity symbols
 *Lists the ORF title and position of motif
 *Outputs a list of counts to stdout.
 */
import java.io.*;
import java.util.*;
import java.util.regex.*;
import org.biojava.bio.*;
import org.biojava.bio.seq.*;
import org.biojava.bio.seq.io.*;
import org.biojava.bio.symbol.*;
public class MotifLister{
  public MotifLister(String type, String inputFile, String target, String placement)throws Exception{
    System.out.println("MotifLister is searching file " + inputFile + " for the motif '" + target +
                        "' in frame " + placement + ".");
    try{
      if(type.equalsIgnoreCase("dna")){
        motif = DNATools.createDNA(target);
      }else if(type.equalsIgnoreCase("rna")){
        motif = RNATools.createRNA(target);
      }else{
        motif = ProteinTools.createProtein(target);
      }
    }
    catch(BioError e){
      System.out.println("Error!!  Data type must match type of motif.");
      System.out.println("Specifically, " + target + " is not " + type);
      System.exit(0);
    }
    Pattern p = Pattern.compile( MotifTools.createRegex(motif) );
    frame = Integer.parseInt(placement);
    if(frame < 0 || frame > 3){
      System.out.println("Only frames 0 through 3 are alloweds");
      System.out.println("frame zero searches all frames.");
      System.exit(0);
    }
    count = 0;
    //read the file
    //input
    FileInputStream fis = new FileInputStream(inputFile);
    InputStreamReader isr = new InputStreamReader(fis);
    BufferedReader input = new BufferedReader(isr);
    try{
      if(type.equalsIgnoreCase("dna")){
        si = SeqIOTools.readFastaDNA(input);
      }else if(type.equalsIgnoreCase("rna")){
        si = SeqIOTools.readFastaRNA(input);
      }else{
        si = SeqIOTools.readFastaProtein(input);
      }
      while (si.hasNext()){
        Sequence seq = si.nextSequence();
        Matcher matcher = p.matcher(seq.seqString());
        int start = 0;
        while(matcher.find(start)) {
          start = matcher.start();
          int end = matcher.end();
          int result = (start % 3) + 1;
          if(result == frame || frame == 0){
            System.out.println(seq.getName() + " : " + "[" + (start + 1) + "," + (end) + "]");
            count++;
          }
          start++;
        }
      }
      input.close(); //close the file
      System.out.println("Total Hits = " + count);
    }
    catch(BioException e){
      System.out.println(inputFile + " is not a " + type + " file.");
      System.out.println(e);
    }
  }
  public static void main(String[] args)throws Exception{
    if (args.length < 4) {
      System.err.println(" Usage: >java -jar MotifLister.jar type fastaFile motif frame" +
                         "\n Ex: >java -jar MotifLister.jar dna eColi.fasta AAAAAAG 3 > output.txt" +
                         "\n would search for A AAA AAG in the third frame in dna file eColi.fasta" +
                         "\n and print the results to file output.txt." +
                         "\n 'type' can be dna, rna, or protein." +
                         "\n 'frame' can be integers 0 through 3." +
                         "\n 0 counts any instance of the motif." +
                         "\n 1, 2, 3 counts only instances of the motif in the specified frame." +
                         "\n Capture output with redirection operator '>'.");
    }else{
      MotifLister ML = new MotifLister(args[0], args[1], args[2], args[3]);
    }
  }
  private SymbolList motif;
  private int frame;
  private int count;
  private SequenceIterator si;
}


---------------------------------
Do you Yahoo!?
Yahoo! Search - Find what you’re looking for faster.
From mark.schreiber at group.novartis.com  Sun Mar  7 20:00:34 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Sun Mar  7 20:03:52 2004
Subject: [Biojava-l] Reading frames and amino acids
Message-ID: <OF7D84F3FC.6B2F584F-ON48256E51.00057C81-48256E51.00053837@EU.novartis.net>

Thanks -

Give me a few days and I'll put this up. If I don't please nag me :)

- Mark





facemann <off2w0rk@yahoo.com>
Sent by: biojava-l-bounces@portal.open-bio.org
03/06/2004 11:09 AM

 
        To:     biojava-l@biojava.org
        cc: 
        Subject:        [Biojava-l] Reading frames and amino acids


Here is a small contribution.  I use it to find simple motifs. Feel free 
to edit or scrap.
 
/**
 *MotifLister.java
 *Andy Hammer
 *08 Aug 2003
 *Lists all instances of a motif in specified (dna\rna\protein) fasta 
file.
 *The motif can contain Ambiguity symbols
 *Lists the ORF title and position of motif
 *Outputs a list of counts to stdout.
 */
import java.io.*;
import java.util.*;
import java.util.regex.*;
import org.biojava.bio.*;
import org.biojava.bio.seq.*;
import org.biojava.bio.seq.io.*;
import org.biojava.bio.symbol.*;
public class MotifLister{
  public MotifLister(String type, String inputFile, String target, String 
placement)throws Exception{
    System.out.println("MotifLister is searching file " + inputFile + " 
for the motif '" + target +
                        "' in frame " + placement + ".");
    try{
      if(type.equalsIgnoreCase("dna")){
        motif = DNATools.createDNA(target);
      }else if(type.equalsIgnoreCase("rna")){
        motif = RNATools.createRNA(target);
      }else{
        motif = ProteinTools.createProtein(target);
      }
    }
    catch(BioError e){
      System.out.println("Error!!  Data type must match type of motif.");
      System.out.println("Specifically, " + target + " is not " + type);
      System.exit(0);
    }
    Pattern p = Pattern.compile( MotifTools.createRegex(motif) );
    frame = Integer.parseInt(placement);
    if(frame < 0 || frame > 3){
      System.out.println("Only frames 0 through 3 are alloweds");
      System.out.println("frame zero searches all frames.");
      System.exit(0);
    }
    count = 0;
    //read the file
    //input
    FileInputStream fis = new FileInputStream(inputFile);
    InputStreamReader isr = new InputStreamReader(fis);
    BufferedReader input = new BufferedReader(isr);
    try{
      if(type.equalsIgnoreCase("dna")){
        si = SeqIOTools.readFastaDNA(input);
      }else if(type.equalsIgnoreCase("rna")){
        si = SeqIOTools.readFastaRNA(input);
      }else{
        si = SeqIOTools.readFastaProtein(input);
      }
      while (si.hasNext()){
        Sequence seq = si.nextSequence();
        Matcher matcher = p.matcher(seq.seqString());
        int start = 0;
        while(matcher.find(start)) {
          start = matcher.start();
          int end = matcher.end();
          int result = (start % 3) + 1;
          if(result == frame || frame == 0){
            System.out.println(seq.getName() + " : " + "[" + (start + 1) + 
"," + (end) + "]");
            count++;
          }
          start++;
        }
      }
      input.close(); //close the file
      System.out.println("Total Hits = " + count);
    }
    catch(BioException e){
      System.out.println(inputFile + " is not a " + type + " file.");
      System.out.println(e);
    }
  }
  public static void main(String[] args)throws Exception{
    if (args.length < 4) {
      System.err.println(" Usage: >java -jar MotifLister.jar type 
fastaFile motif frame" +
                         "\n Ex: >java -jar MotifLister.jar dna 
eColi.fasta AAAAAAG 3 > output.txt" +
                         "\n would search for A AAA AAG in the third frame 
in dna file eColi.fasta" +
                         "\n and print the results to file output.txt." +
                         "\n 'type' can be dna, rna, or protein." +
                         "\n 'frame' can be integers 0 through 3." +
                         "\n 0 counts any instance of the motif." +
                         "\n 1, 2, 3 counts only instances of the motif in 
the specified frame." +
                         "\n Capture output with redirection operator 
'>'.");
    }else{
      MotifLister ML = new MotifLister(args[0], args[1], args[2], 
args[3]);
    }
  }
  private SymbolList motif;
  private int frame;
  private int count;
  private SequenceIterator si;
}


---------------------------------
Do you Yahoo!?
Yahoo! Search - Find what you're looking for 
faster._______________________________________________
Biojava-l mailing list  -  Biojava-l@biojava.org
http://biojava.org/mailman/listinfo/biojava-l




From Gera.Jellema at wur.nl  Mon Mar  8 10:50:40 2004
From: Gera.Jellema at wur.nl (Jellema, Gera)
Date: Mon Mar  8 10:56:25 2004
Subject: [Biojava-l] Reading frames and amino acids
Message-ID: <EF1BA7172DC75F4182B8E8A5B08EF6400E5ED9@salte0008.wurnet.nl>

I tried to compile the program Mark Schreiber wrote but the only thing I get is:
Hex.java:13: illegal character: \154
  /**
 ^
Hex.java:16: illegal character: \154
  public static void help() {
 ^
Hex.java:17: illegal character: \154
  System.out.println(
 ^
Hex.java:17: illegal character: \154
  System.out.println(
  ^
Hex.java:17: illegal character: \154
  System.out.println(
   ^
Hex.java:18: illegal character: \154
  "usage: java utils.Hex <file> <format eg fasta> <dna|rna>");
 ^
Hex.java:18: illegal character: \154
  "usage: java utils.Hex <file> <format eg fasta> <dna|rna>");
  ^
In total I get 100 of these error messages, and I don't understand them. Can you help me?
Gera


-----Oorspronkelijk bericht-----
Van:	mark.schreiber@group.novartis.com [mailto:mark.schreiber@group.novartis.com]
Verzonden:	vr 5-3-2004 10:35
Aan:	Dan Bolser
CC:	biojava-l@portal.open-bio.org; biojava-l-bounces@portal.open-bio.org; Jellema, Gera
Onderwerp:	Re: [Biojava-l] Reading frames and amino acids

That is the intention of the 'biojava in anger' site. I just haven't had 
time to add lots of little scripts to it. I will gratefully take 
donations.

- Mark






Dan Bolser <dmb@mrc-dunn.cam.ac.uk>
03/05/2004 05:29 PM

 
        To:     Mark Schreiber/GP/Novartis@PH
        cc:     "Jellema, Gera" <Gera.Jellema@wur.nl>, biojava-l@portal.open-bio.org, 
biojava-l-bounces@portal.open-bio.org
        Subject:        Re: [Biojava-l] Reading frames and amino acids




I have seen 'biojava in anger', but is their a 'biojava cookbook'?

It strikes me that lots of small scripts like this could be useful for
lots of people if archived properly.

Does the cookbook exist?

Cheers, Dan.

On Fri, 5 Mar 2004 mark.schreiber@group.novartis.com wrote:

> Hi Gera -
> 
> The code below seems to work for me. Please note that I have not tested 
it 
> thoroughly so you might want to eyeball a few results to make sure they 
> are sensible.
> 
> You could probably improve it by making some of the 'in line' code into 
> methods etc.
> 
> - Mark
> 
> import java.io.*;
> 
> import org.biojava.bio.*;
> import org.biojava.bio.seq.*;
> import org.biojava.bio.seq.io.*;
> import org.biojava.bio.symbol.*;
> 
> /**
>  * <p>Program to six-frame translate a nucleotide sequence</p>
>  */
> 
> public class Hex {
>   /**
>    * Call this to get usage info, program terminates after call.
>    */
>   public static void help() {
>     System.out.println(
>         "usage: java utils.Hex <file> <format eg fasta> <dna|rna>");
>     System.exit( -1);
>   }
> 
>   public static void main(String[] args) throws Exception{
>     if (args.length != 3) {
>       help();
>     }
> 
>     BufferedReader br = null;
>     String format = args[1];
>     String alpha = args[2];
> 
>     try {
>       br = new BufferedReader(new FileReader(args[0]));
> 
>       SequenceIterator seqi = 
>           (SequenceIterator)SeqIOTools.fileToBiojava(format, alpha, br);
> 
>       //for each sequence
>       while(seqi.hasNext()){
>         Sequence seq = seqi.nextSequence();
> 
>         //for each frame
>         for (int i = 0; i < 3; i++) {
>           SymbolList prot;
>           Sequence trans;
> 
>           //take the reading frame
>           SymbolList syms = seq.subList(
>                 i+1,
>                 seq.length() - (seq.length() - i)%3);
> 
> 
>           //if it is DNA transcribe it to RNA 
>           if(syms.getAlphabet() == DNATools.getDNA()){
>             syms = RNATools.transcribe(syms);
>           }
> 
>           //output forward translation to STDOUT
>           prot = RNATools.translate(syms);
>           trans = SequenceTools.createSequence(prot, "", 
>                                                seq.getName()+
>                                                "TranslationFrame: +"+i,
>  Annotation.EMPTY_ANNOTATION);
>           SeqIOTools.writeFasta(System.out, trans);
> 
>           //output reverse frame translation to STDOUT
>           syms = RNATools.reverseComplement(syms);
>           prot = RNATools.translate(syms);
>           trans = SequenceTools.createSequence(prot, "",
>                                                seq.getName() +
>                                                "TranslationFrame: -" + 
i,
>  Annotation.EMPTY_ANNOTATION);
>           SeqIOTools.writeFasta(System.out, trans);
>         }
>       }
>     }
>     finally {
>       if(br != null){
>         br.close();
>       }
>     }
>   }
> }
> 
> 
> 
> 
> 
> "Jellema, Gera" <Gera.Jellema@wur.nl>
> Sent by: biojava-l-bounces@portal.open-bio.org
> 03/04/2004 09:30 PM
> 
> 
>         To:     <biojava-l@portal.open-bio.org>
>         cc: 
>         Subject:        [Biojava-l] Reading frames and amino acids
> 
> 
> Hi,
> I'm new to biojava so I don't know if this has already been asked. I 
have 
> a genome sequence and I want to have it in the 6 reading frames, and 
then 
> per reading frame translated into amino acids so I can look for 
proteins. 
> I don't know how I have to do it.
> Thanks,
> Gera
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 
> [ Attachment ''WINMAIL.DAT'' removed by Mark Schreiber ]
> 
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 







From smh1008 at cus.cam.ac.uk  Mon Mar  8 11:37:16 2004
From: smh1008 at cus.cam.ac.uk (David Huen)
Date: Mon Mar  8 11:43:18 2004
Subject: [Biojava-l] Reading frames and amino acids
In-Reply-To: <EF1BA7172DC75F4182B8E8A5B08EF6400E5ED9@salte0008.wurnet.nl>
References: <EF1BA7172DC75F4182B8E8A5B08EF6400E5ED9@salte0008.wurnet.nl>
Message-ID: <200403081637.16804.smh1008@cus.cam.ac.uk>

On Monday 08 Mar 2004 3:50 pm, Jellema, Gera wrote:
> I tried to compile the program Mark Schreiber wrote but the only thing I
> get is: Hex.java:13: illegal character: \154
>   /**
>  ^
> Hex.java:16: illegal character: \154
>   public static void help() {
>  ^
> Hex.java:17: illegal character: \154
>   System.out.println(

Is there any possibility that you are using an editor that inserts invisible 
characters and/or some usual end-of-line character?  You need to save in 
plain text.  Also I don't know how well javac will tolerate DOS <CR><LF> 
line ends.

Regards,
David Huen

From mark.schreiber at group.novartis.com  Wed Mar 10 20:08:32 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Wed Mar 10 20:14:19 2004
Subject: [Biojava-l] BJIA updates
Message-ID: <OF5F21AA5C.A7040BAB-ON48256E54.0005B8F5-48256E54.0006465F@EU.novartis.net>

Hi All -

I have updated the BioJava in Anger site 
(http://www.biojava.org/docs/bj_in_anger/) adding some new tutorials and a 
bug fix.

Specifically, there are new tutorials on:

* Calculating mass and pI of a peptide 
(http://www.biojava.org/docs/bj_in_anger/calcmass.html)
* Creating a regular expression from a sequence motif 
(http://www.biojava.org/docs/bj_in_anger/regex.html) thanks to Andy Hammer 
for this sample!
* Performing a six frame translation 
(http://www.biojava.org/docs/bj_in_anger/sixframetranslate.html)

I have also fixed the bug in the view sequence demo 
(http://www.biojava.org/docs/bj_in_anger/NameChange.htm). This now uses 
the recommended API for generating ViewSequences.

Enjoy

- Mark

Mark Schreiber
Principal Scientist (Bioinformatics)

Novartis Institute for Tropical Diseases (NITD)
1 Science Park Road
#04-14 The Capricorn, Science Park II
Singapore 117528

phone +65 6722 2973
fax  +65 6722 2910

From dag at sonsorol.org  Wed Mar 10 21:30:13 2004
From: dag at sonsorol.org (Chris Dagdigian)
Date: Wed Mar 10 21:35:53 2004
Subject: [Biojava-l] O|B|F mail update -- making progress on anti-spam
 issues with our mailing lists
Message-ID: <404FCF35.5010705@sonsorol.org>


Hi folks,

Apologies for the cross-posting but I just wanted to give our list 
members and admins an update on some new anti-spam measures we have 
(re)enabled. Good news to report basically...

The most annoying spams recently have been the simple plain text 
messages without any HTML, attachments or mime-encoding that just slip 
right by our filters.  Some lists have been forced to switch over to 
"only members can post" while other lists (like bioperl) have 
consistantly voted to stay as open as possible.

I'll update you on our current efforts as well as a new effort that is 
about 24 hours old but already working really well so far.

Until yesterday we had three main lines of defense against spam:

1. The mailserver itself (rejects mail from nonexistant domains, etc.)

2. The sendmail Mail::Milter extention (MIMEDefang+SpamAssassin are used 
to scan all incoming messages. Anything that scores higher than 8.0 is 
simply discarded automatically. MIMEDefang also strips dangerous 
attachments like .exe and .pif)

3. Our mailing list moderation queue (emails with attachments, odd MIME 
encodings and spamassassin scores from 0.0 - 7.9 are held in a moderator 
queue for a human to make an accept/discard decision)

Here are some stats on how this system worked over the past few days:

  o 138 attempts to relay mail through our server blocked
  o 192 emails blocked due to forged or unresolvable sender domain
  o 577 emails discarded automatically by SpamAssassin+MIMEDefang

This system worked *ok* but put a lot of work onto the shoulders of our 
list admins who constantly had to weed out the spam caught up in the 
mailing list moderator system.

Yesterday I brought online another system that seems to be already 
working really well. It catches spam before we even accept it on our 
server which makes the load easier on both our scanning software and our 
  human list moderators.

The system is the RBL+ blackhole list from http://www.mail-abuse.org and 
the way it works is that we now query (via DNS) the RBL+ database each 
time someone connects to our mail server. If the RBL check against the 
sender IP address comes back as "positive" we reject the incoming email.

RBL+ is a combination of four constantly updated databases:

  1. RBL -- IP addresses of known, documented spammers and spam machines
  2. RSS -- IP addresses of documented/tested unsecured email relays
  3. OPS -- IP addresses of documented open proxy servers w/ spam history
  3. DUL -- IP addresses belonging to ISP dialup and DHCP customers

We have already blocked 137 email attempts in the last 24 hours from 
machines that were listed in one or more of the RBL databases.

It is too soon to tell but if the RBL+ system plus our existing 
anti-spam measures work well enough we may be in a position where our 
"closed" mailing lists could revert back to being 'anyone can post'.

Feedback appreciated. Especially if you get a "reject" message from us 
saying that you are listed in the RBL+ blackhole database!


Regards,
Chris
O|B|F




From DMGoodstein at lbl.gov  Thu Mar 11 15:37:31 2004
From: DMGoodstein at lbl.gov (DMGoodstein@lbl.gov)
Date: Thu Mar 11 15:43:09 2004
Subject: [Biojava-l] BlastXMLParserFacade
Message-ID: <43dfea43ae54.43ae5443dfea@lbl.gov>

I was wondering if anyone has successfully gotten
BlastXMLParserFacade to work on an xml style NCBI
blast output file (version 2.2.3)? I'm getting null
pointer exceptions deep within the crimson parser
implementation classes, but I now i'm correctly
passing in the NCBI output file, since the dtd is
getting read.

--David Goodstein
  Joint Genome Institute

/usr/java/j2sdk1.4.1_02/bin/java2/bin/java
-classpath .:./biojava-1.3.1
.1.jar BlastParser fred.xml
a/j2sdk1.4.1_02/bin/javac -classpath .:./biojava-1.3.
java.lang.NullPointerException
	at
org.apache.crimson.parser.Parser2.parseInternal(Parser2.java:524)
	at
org.apache.crimson.parser.Parser2.parse(Parser2.java:305)
	at
org.apache.crimson.parser.XMLReaderImpl.parse(XMLReaderImpl.java:442)
	at
org.biojava.bio.program.sax.blastxml.BlastXMLParserFacade.parse(BlastXMLParserFacade.java:172)
	at BlastParser.main(BlastParser.java:44)
1.jar BlastParser
./fred.xmlava/j2sdk1.4.1_02/bin/java -classpath
.:./biojava-1.3.1
java.lang.NullPointerException
	at
org.apache.crimson.parser.Parser2.parseInternal(Parser2.java:524)
	at
org.apache.crimson.parser.Parser2.parse(Parser2.java:305)
	at
org.apache.crimson.parser.XMLReaderImpl.parse(XMLReaderImpl.java:442)
	at
org.biojava.bio.program.sax.blastxml.BlastXMLParserFacade.parse(BlastXMLParserFacade.java:172)
	at BlastParser.main(BlastParser.java:44)

From sacoca at MCB.McGill.CA  Thu Mar 11 17:11:02 2004
From: sacoca at MCB.McGill.CA (sacoca@MCB.McGill.CA)
Date: Thu Mar 11 17:16:37 2004
Subject: [Biojava-l] Parameter Settings in BaumWelchTraining
Message-ID: <3505.66.131.112.37.1079043062.squirrel@mail.MCB.McGill.CA>

Hi all. I'm trying to optimize the transition states probabilities for my
HMM. I already have set them to values which I think are pretty good.
Since I know the Baum Welch can only help with the scores and optimize
them up to a local maxima I thought of using the parameters I calculated
as a starting point. The problem is that I don't know how!
I followed the example in biojava:

....
//train the model to have uniform parameters
    ModelTrainer mt = new SimpleModelTrainer();
    //register the model to train
    mt.registerModel(hmm);

I want to use the values already set in my hmm  as the starting parameters
in the BaumWelch.  I don't want to use the uniform distribution as
indicated below!

    //as no other counts are being used the null weight will cause   
everything to be uniform
    mt.setNullModelWeight(1.0);
    mt.train();

I tried adding counts and looking up examples on the net but ended up more
confused than I started. How do I use the addCounts to make this work!

Stephane Acoca
Master's Student
McGill Center for Bioinformatics

From mark.schreiber at group.novartis.com  Thu Mar 11 20:58:52 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Thu Mar 11 21:04:33 2004
Subject: [Biojava-l] Parameter Settings in BaumWelchTraining
Message-ID: <OFF37EB8C6.5D68B562-ON48256E55.00098D98-48256E55.000AE1FB@EU.novartis.net>

Hi Stephane -

Within EmissionState you can set a Distribution that contains emission 
probabilities for the Symbols states emission alphabet using the 
setDistribution method. This Distribution will be your predetermined 
weights.

To set the transition probabilities you can use the setWeights(State 
source, Distribution weights). The source is the state you are 
transitioning from and the weights is the probability of transitioning to 
any State that the source connects too. Because States implement Symbol 
you can put them in a Distribution.

To make a Distribution of States that state 'a' could connect to use the 
following pseudo code:

State a;
Model m;
FiniteAlphabet endPoints;

endPoints = m.transitionsFrom(a);
Distribution d = 
DistributionFactory.DEFAULT.createDistribution(endPoints);

//You can then train d or set it's weights and put it back in the model 
with

m.setWeights(a, d);

Mark Schreiber
Principal Scientist (Bioinformatics)

Novartis Institute for Tropical Diseases (NITD)
1 Science Park Road
#04-14 The Capricorn, Science Park II
Singapore 117528

phone +65 6722 2973
fax  +65 6722 2910





sacoca@mcb.mcgill.ca
Sent by: biojava-l-bounces@portal.open-bio.org
03/12/2004 06:11 AM

 
        To:     "Biojava Mailing List" <biojava-l@biojava.org>
        cc: 
        Subject:        [Biojava-l] Parameter Settings in BaumWelchTraining


Hi all. I'm trying to optimize the transition states probabilities for my
HMM. I already have set them to values which I think are pretty good.
Since I know the Baum Welch can only help with the scores and optimize
them up to a local maxima I thought of using the parameters I calculated
as a starting point. The problem is that I don't know how!
I followed the example in biojava:

....
//train the model to have uniform parameters
    ModelTrainer mt = new SimpleModelTrainer();
    //register the model to train
    mt.registerModel(hmm);

I want to use the values already set in my hmm  as the starting parameters
in the BaumWelch.  I don't want to use the uniform distribution as
indicated below!

    //as no other counts are being used the null weight will cause 
everything to be uniform
    mt.setNullModelWeight(1.0);
    mt.train();

I tried adding counts and looking up examples on the net but ended up more
confused than I started. How do I use the addCounts to make this work!

Stephane Acoca
Master's Student
McGill Center for Bioinformatics

_______________________________________________
Biojava-l mailing list  -  Biojava-l@biojava.org
http://biojava.org/mailman/listinfo/biojava-l



From sacoca at MCB.McGill.CA  Fri Mar 12 00:28:05 2004
From: sacoca at MCB.McGill.CA (sacoca@MCB.McGill.CA)
Date: Fri Mar 12 00:33:38 2004
Subject: [Biojava-l] Parameter Settings in BaumWelchTraining
In-Reply-To: <OFF37EB8C6.5D68B562-ON48256E55.00098D98-48256E55.000AE1FB@EU.novartis
	.net>
References: <OFF37EB8C6.5D68B562-ON48256E55.00098D98-48256E55.000AE1FB@EU.novartis.net>
Message-ID: <4182.66.131.112.37.1079069285.squirrel@mail.MCB.McGill.CA>

> Hi Stephane -
>
> Within EmissionState you can set a Distribution that contains emission
> probabilities for the Symbols states emission alphabet using the
> setDistribution method. This Distribution will be your predetermined
> weights.
>
> To set the transition probabilities you can use the setWeights(State
> source, Distribution weights). The source is the state you are
> transitioning from and the weights is the probability of transitioning to
> any State that the source connects too. Because States implement Symbol
> you can put them in a Distribution.
>
> To make a Distribution of States that state 'a' could connect to use the
> following pseudo code:
>
> State a;
> Model m;
> FiniteAlphabet endPoints;
>
> endPoints = m.transitionsFrom(a);
> Distribution d =
> DistributionFactory.DEFAULT.createDistribution(endPoints);
>
> //You can then train d or set it's weights and put it back in the model
> with
>
> m.setWeights(a, d);
>
> Mark Schreiber
> Principal Scientist (Bioinformatics)
>
> Novartis Institute for Tropical Diseases (NITD)
> 1 Science Park Road
> #04-14 The Capricorn, Science Park II
> Singapore 117528
>
> phone +65 6722 2973
> fax  +65 6722 2910
>
>
>
>
>
> sacoca@mcb.mcgill.ca
> Sent by: biojava-l-bounces@portal.open-bio.org
> 03/12/2004 06:11 AM
>
>
>         To:     "Biojava Mailing List" <biojava-l@biojava.org>
>         cc:
>         Subject:        [Biojava-l] Parameter Settings in
> BaumWelchTraining
>
>
> Hi all. I'm trying to optimize the transition states probabilities for my
> HMM. I already have set them to values which I think are pretty good.
> Since I know the Baum Welch can only help with the scores and optimize
> them up to a local maxima I thought of using the parameters I calculated
> as a starting point. The problem is that I don't know how!
> I followed the example in biojava:
>
> ....
> //train the model to have uniform parameters
>     ModelTrainer mt = new SimpleModelTrainer();
>     //register the model to train
>     mt.registerModel(hmm);
>
> I want to use the values already set in my hmm  as the starting parameters
> in the BaumWelch.  I don't want to use the uniform distribution as
> indicated below!
>
>     //as no other counts are being used the null weight will cause
> everything to be uniform
>     mt.setNullModelWeight(1.0);
>     mt.train();
>
> I tried adding counts and looking up examples on the net but ended up more
> confused than I started. How do I use the addCounts to make this work!
>
> Stephane Acoca
> Master's Student
> McGill Center for Bioinformatics
>
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
>
>
>

From sacoca at MCB.McGill.CA  Fri Mar 12 00:30:16 2004
From: sacoca at MCB.McGill.CA (sacoca@MCB.McGill.CA)
Date: Fri Mar 12 00:35:49 2004
Subject: [Biojava-l] Parameter Settings in BaumWelchTraining]
Message-ID: <4195.66.131.112.37.1079069416.squirrel@mail.MCB.McGill.CA>

Sorry for the previous error.
---------------------------- Original Message ----------------------------
Subject: Re: [Biojava-l] Parameter Settings in BaumWelchTraining
From:    sacoca@MCB.McGill.CA
Date:    Fri, March 12, 2004 12:27 am
To:      mark.schreiber@group.novartis.com
--------------------------------------------------------------------------

Here is the code I have for the training. Using what you told me below, I
can retreive all of the weights that I calculated manually for the hmm
(distributions for the transitions and distributions for the alphabet of  
each state). What I do not understand is how to use this information and
the sequences stored in a file to run the BaumWelchAlgorithm and then
retreive the optimized values calculated by the algorithm to set them back
into my HMM.

//Retreive the alphabet of all states
FiniteAlphabet SA = hmm.stateAlphabet();
Iterator i = SA.iterator();

SimpleModelTrainer MT = new SimpleModelTrainer();
MT.registerModel(hmm);

//go through each state
while(i.hasNext())
{Symbol Currentstate = (Symbol)i.next();

 //Retreive the distribution of all transitions from the current state
FiniteAlphabet From = hmm.transitionsFrom((State)Currentstate);
 Distribution d = hmm.getWeights((State)Currentstate);
 Iterator i2 = From.iterator();

 //go through it and look at all the weights for each of the transitions
while(i2.hasNext())
    {Symbol s = (Symbol)i2.next();
     System.out.println("From state "+Currentstate.getName()+
                        "To State   "+s.getName()+
                         "Weight     "+d.getWeight(s));}

 //get the distribution for the alphabet of the current state
 Distribution d2 =((EmissionState)Currentstate).getDistribution();
FiniteAlphabet IN = (FiniteAlphabet)hmm.emissionAlphabet();
 Iterator i3 = IN.iterator();
 //you can go through it the same way as above using a while loop
*****************************************************************
This is what I don't understand!!!!
*****************************************************************
here, we have a set of training sequences stored in a file in fasta format
that i'd like to use with the BaumWelch algorithm to optimize the
transition distributions mentionned above.

//This is the file with all the training sequences
BufferedInputStream is = new BufferedInputStream(new
FileInputStream("z:/Sequences.faa"));

//Load the file with the SequenceDB class
SequenceDB DB = SeqIOTools.readFasta(is, ProtAlphabet);

//use 100 cycles as the stop criteria
StoppingCriteria stopper = new StoppingCriteria()
     {public boolean isTrainingComplete(TrainingAlgorithm ta)
       {return (ta.getCycle() > 100);}};

*****************************************
This part is what I am clueless about
*****************************************
//How do I optimize my hmm with the BaumWelch algorithm and retreive //the
optimized values ? How do I train the distribution above with //the baum
welch and the sequences that I have ?
DP dp= DPFactory.DEFAULT.createDP(hmm);
BaumWelchTrainer bwt = new BaumWelchTrainer(dp);
}

PS : I do not know why you are helping all of us here but thank you. It
makes Biojava a lot easier to deal with.

Steve

> Hi Stephane -
>
> Within EmissionState you can set a Distribution that contains emission
probabilities for the Symbols states emission alphabet using the
setDistribution method. This Distribution will be your predetermined
weights.
>
> To set the transition probabilities you can use the setWeights(State
source, Distribution weights). The source is the state you are
> transitioning from and the weights is the probability of transitioning
to any State that the source connects too. Because States implement
Symbol you can put them in a Distribution.
>
> To make a Distribution of States that state 'a' could connect to use the
following pseudo code:
>
> State a;
> Model m;
> FiniteAlphabet endPoints;
>
> endPoints = m.transitionsFrom(a);
> Distribution d =
> DistributionFactory.DEFAULT.createDistribution(endPoints);
>
> //You can then train d or set it's weights and put it back in the model
with
>
> m.setWeights(a, d);
>
> Mark Schreiber
> Principal Scientist (Bioinformatics)
>
> Novartis Institute for Tropical Diseases (NITD)
> 1 Science Park Road
> #04-14 The Capricorn, Science Park II
> Singapore 117528
>
> phone +65 6722 2973
> fax  +65 6722 2910
>
>
>
>
>
> sacoca@mcb.mcgill.ca
> Sent by: biojava-l-bounces@portal.open-bio.org
> 03/12/2004 06:11 AM
>
>
>         To:     "Biojava Mailing List" <biojava-l@biojava.org>
>         cc:
>         Subject:        [Biojava-l] Parameter Settings in
> BaumWelchTraining
>
>
> Hi all. I'm trying to optimize the transition states probabilities for
my HMM. I already have set them to values which I think are pretty good.
Since I know the Baum Welch can only help with the scores and optimize
them up to a local maxima I thought of using the parameters I calculated
as a starting point. The problem is that I don't know how!
> I followed the example in biojava:
>
> ....
> //train the model to have uniform parameters
>     ModelTrainer mt = new SimpleModelTrainer();
>     //register the model to train
>     mt.registerModel(hmm);
>
> I want to use the values already set in my hmm  as the starting
parameters in the BaumWelch.  I don't want to use the uniform
distribution as indicated below!
>
>     //as no other counts are being used the null weight will cause
> everything to be uniform
>     mt.setNullModelWeight(1.0);
>     mt.train();
>
> I tried adding counts and looking up examples on the net but ended up
more confused than I started. How do I use the addCounts to make this
work!
>
> Stephane Acoca
> Master's Student
> McGill Center for Bioinformatics
>
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
>
>
>



From mark.schreiber at group.novartis.com  Fri Mar 12 01:00:43 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Fri Mar 12 01:06:22 2004
Subject: [Biojava-l] Parameter Settings in BaumWelchTraining]
Message-ID: <OF092DE640.8908A512-ON48256E55.00205201-48256E55.0021064A@EU.novartis.net>

When you call the train() method of the BaumWelchTrainer you supply it 
with a SequenceDB. The sequences from this DB are used to optimize the 
weights of the model.

However, I have a bad feeling that when you train your model with the 
BaumWelchTrainer your previously set counts will be ignored and 
overwritten. You could check by looking into AbstractModelTrainer.train() 
(which is what the BaumWelchTrainer extends). You could also run some 
tests to see if using a pre-trained model makes any difference to the 
final outcome. Does anyone more expert than me on the DP package (ie most 
people) know if the counts are overwritten?

- Mark





sacoca@mcb.mcgill.ca
Sent by: biojava-l-bounces@portal.open-bio.org
03/12/2004 01:30 PM

 
        To:     sacoca@mcb.mcgill.ca
        cc:     Biojava Mailing List <biojava-l@biojava.org>
        Subject:        Re: [Biojava-l] Parameter Settings in BaumWelchTraining]


Sorry for the previous error.
---------------------------- Original Message ----------------------------
Subject: Re: [Biojava-l] Parameter Settings in BaumWelchTraining
From:    sacoca@MCB.McGill.CA
Date:    Fri, March 12, 2004 12:27 am
To:      mark.schreiber@group.novartis.com
--------------------------------------------------------------------------

Here is the code I have for the training. Using what you told me below, I
can retreive all of the weights that I calculated manually for the hmm
(distributions for the transitions and distributions for the alphabet of 
each state). What I do not understand is how to use this information and
the sequences stored in a file to run the BaumWelchAlgorithm and then
retreive the optimized values calculated by the algorithm to set them back
into my HMM.

//Retreive the alphabet of all states
FiniteAlphabet SA = hmm.stateAlphabet();
Iterator i = SA.iterator();

SimpleModelTrainer MT = new SimpleModelTrainer();
MT.registerModel(hmm);

//go through each state
while(i.hasNext())
{Symbol Currentstate = (Symbol)i.next();

 //Retreive the distribution of all transitions from the current state
FiniteAlphabet From = hmm.transitionsFrom((State)Currentstate);
 Distribution d = hmm.getWeights((State)Currentstate);
 Iterator i2 = From.iterator();

 //go through it and look at all the weights for each of the transitions
while(i2.hasNext())
    {Symbol s = (Symbol)i2.next();
     System.out.println("From state "+Currentstate.getName()+
                        "To State   "+s.getName()+
                         "Weight     "+d.getWeight(s));}

 //get the distribution for the alphabet of the current state
 Distribution d2 =((EmissionState)Currentstate).getDistribution();
FiniteAlphabet IN = (FiniteAlphabet)hmm.emissionAlphabet();
 Iterator i3 = IN.iterator();
 //you can go through it the same way as above using a while loop
*****************************************************************
This is what I don't understand!!!!
*****************************************************************
here, we have a set of training sequences stored in a file in fasta format
that i'd like to use with the BaumWelch algorithm to optimize the
transition distributions mentionned above.

//This is the file with all the training sequences
BufferedInputStream is = new BufferedInputStream(new
FileInputStream("z:/Sequences.faa"));

//Load the file with the SequenceDB class
SequenceDB DB = SeqIOTools.readFasta(is, ProtAlphabet);

//use 100 cycles as the stop criteria
StoppingCriteria stopper = new StoppingCriteria()
     {public boolean isTrainingComplete(TrainingAlgorithm ta)
       {return (ta.getCycle() > 100);}};

*****************************************
This part is what I am clueless about
*****************************************
//How do I optimize my hmm with the BaumWelch algorithm and retreive //the
optimized values ? How do I train the distribution above with //the baum
welch and the sequences that I have ?
DP dp= DPFactory.DEFAULT.createDP(hmm);
BaumWelchTrainer bwt = new BaumWelchTrainer(dp);
}

PS : I do not know why you are helping all of us here but thank you. It
makes Biojava a lot easier to deal with.

Steve

> Hi Stephane -
>
> Within EmissionState you can set a Distribution that contains emission
probabilities for the Symbols states emission alphabet using the
setDistribution method. This Distribution will be your predetermined
weights.
>
> To set the transition probabilities you can use the setWeights(State
source, Distribution weights). The source is the state you are
> transitioning from and the weights is the probability of transitioning
to any State that the source connects too. Because States implement
Symbol you can put them in a Distribution.
>
> To make a Distribution of States that state 'a' could connect to use the
following pseudo code:
>
> State a;
> Model m;
> FiniteAlphabet endPoints;
>
> endPoints = m.transitionsFrom(a);
> Distribution d =
> DistributionFactory.DEFAULT.createDistribution(endPoints);
>
> //You can then train d or set it's weights and put it back in the model
with
>
> m.setWeights(a, d);
>
> Mark Schreiber
> Principal Scientist (Bioinformatics)
>
> Novartis Institute for Tropical Diseases (NITD)
> 1 Science Park Road
> #04-14 The Capricorn, Science Park II
> Singapore 117528
>
> phone +65 6722 2973
> fax  +65 6722 2910
>
>
>
>
>
> sacoca@mcb.mcgill.ca
> Sent by: biojava-l-bounces@portal.open-bio.org
> 03/12/2004 06:11 AM
>
>
>         To:     "Biojava Mailing List" <biojava-l@biojava.org>
>         cc:
>         Subject:        [Biojava-l] Parameter Settings in
> BaumWelchTraining
>
>
> Hi all. I'm trying to optimize the transition states probabilities for
my HMM. I already have set them to values which I think are pretty good.
Since I know the Baum Welch can only help with the scores and optimize
them up to a local maxima I thought of using the parameters I calculated
as a starting point. The problem is that I don't know how!
> I followed the example in biojava:
>
> ....
> //train the model to have uniform parameters
>     ModelTrainer mt = new SimpleModelTrainer();
>     //register the model to train
>     mt.registerModel(hmm);
>
> I want to use the values already set in my hmm  as the starting
parameters in the BaumWelch.  I don't want to use the uniform
distribution as indicated below!
>
>     //as no other counts are being used the null weight will cause
> everything to be uniform
>     mt.setNullModelWeight(1.0);
>     mt.train();
>
> I tried adding counts and looking up examples on the net but ended up
more confused than I started. How do I use the addCounts to make this
work!
>
> Stephane Acoca
> Master's Student
> McGill Center for Bioinformatics
>
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
>
>
>



_______________________________________________
Biojava-l mailing list  -  Biojava-l@biojava.org
http://biojava.org/mailman/listinfo/biojava-l



From dmb at mrc-dunn.cam.ac.uk  Fri Mar 12 03:47:50 2004
From: dmb at mrc-dunn.cam.ac.uk (Dan Bolser)
Date: Fri Mar 12 03:50:11 2004
Subject: [Biojava-l] Parameter Settings in BaumWelchTraining]
In-Reply-To: <OF092DE640.8908A512-ON48256E55.00205201-48256E55.0021064A@EU.novartis.net>
Message-ID: <Pine.LNX.4.21.0403120846470.3749-100000@mail.mrc-dunn.cam.ac.uk>

On Fri, 12 Mar 2004 mark.schreiber@group.novartis.com wrote:

> When you call the train() method of the BaumWelchTrainer you supply it 
> with a SequenceDB. The sequences from this DB are used to optimize the 
> weights of the model.
> 
> However, I have a bad feeling that when you train your model with the 
> BaumWelchTrainer your previously set counts will be ignored and 
> overwritten. You could check by looking into AbstractModelTrainer.train() 
> (which is what the BaumWelchTrainer extends). You could also run some 
> tests to see if using a pre-trained model makes any difference to the 
> final outcome. Does anyone more expert than me on the DP package (ie most 
> people) know if the counts are overwritten?

The idea sounds good either way, so it would be a shame to have to reject
it on the basis of a technicality :)

Cheers

> 
> - Mark
> 
> 
> 
> 
> 
> sacoca@mcb.mcgill.ca
> Sent by: biojava-l-bounces@portal.open-bio.org
> 03/12/2004 01:30 PM
> 
>  
>         To:     sacoca@mcb.mcgill.ca
>         cc:     Biojava Mailing List <biojava-l@biojava.org>
>         Subject:        Re: [Biojava-l] Parameter Settings in BaumWelchTraining]
> 
> 
> Sorry for the previous error.
> ---------------------------- Original Message ----------------------------
> Subject: Re: [Biojava-l] Parameter Settings in BaumWelchTraining
> From:    sacoca@MCB.McGill.CA
> Date:    Fri, March 12, 2004 12:27 am
> To:      mark.schreiber@group.novartis.com
> --------------------------------------------------------------------------
> 
> Here is the code I have for the training. Using what you told me below, I
> can retreive all of the weights that I calculated manually for the hmm
> (distributions for the transitions and distributions for the alphabet of 
> each state). What I do not understand is how to use this information and
> the sequences stored in a file to run the BaumWelchAlgorithm and then
> retreive the optimized values calculated by the algorithm to set them back
> into my HMM.
> 
> //Retreive the alphabet of all states
> FiniteAlphabet SA = hmm.stateAlphabet();
> Iterator i = SA.iterator();
> 
> SimpleModelTrainer MT = new SimpleModelTrainer();
> MT.registerModel(hmm);
> 
> //go through each state
> while(i.hasNext())
> {Symbol Currentstate = (Symbol)i.next();
> 
>  //Retreive the distribution of all transitions from the current state
> FiniteAlphabet From = hmm.transitionsFrom((State)Currentstate);
>  Distribution d = hmm.getWeights((State)Currentstate);
>  Iterator i2 = From.iterator();
> 
>  //go through it and look at all the weights for each of the transitions
> while(i2.hasNext())
>     {Symbol s = (Symbol)i2.next();
>      System.out.println("From state "+Currentstate.getName()+
>                         "To State   "+s.getName()+
>                          "Weight     "+d.getWeight(s));}
> 
>  //get the distribution for the alphabet of the current state
>  Distribution d2 =((EmissionState)Currentstate).getDistribution();
> FiniteAlphabet IN = (FiniteAlphabet)hmm.emissionAlphabet();
>  Iterator i3 = IN.iterator();
>  //you can go through it the same way as above using a while loop
> *****************************************************************
> This is what I don't understand!!!!
> *****************************************************************
> here, we have a set of training sequences stored in a file in fasta format
> that i'd like to use with the BaumWelch algorithm to optimize the
> transition distributions mentionned above.
> 
> //This is the file with all the training sequences
> BufferedInputStream is = new BufferedInputStream(new
> FileInputStream("z:/Sequences.faa"));
> 
> //Load the file with the SequenceDB class
> SequenceDB DB = SeqIOTools.readFasta(is, ProtAlphabet);
> 
> //use 100 cycles as the stop criteria
> StoppingCriteria stopper = new StoppingCriteria()
>      {public boolean isTrainingComplete(TrainingAlgorithm ta)
>        {return (ta.getCycle() > 100);}};
> 
> *****************************************
> This part is what I am clueless about
> *****************************************
> //How do I optimize my hmm with the BaumWelch algorithm and retreive //the
> optimized values ? How do I train the distribution above with //the baum
> welch and the sequences that I have ?
> DP dp= DPFactory.DEFAULT.createDP(hmm);
> BaumWelchTrainer bwt = new BaumWelchTrainer(dp);
> }
> 
> PS : I do not know why you are helping all of us here but thank you. It
> makes Biojava a lot easier to deal with.
> 
> Steve
> 
> > Hi Stephane -
> >
> > Within EmissionState you can set a Distribution that contains emission
> probabilities for the Symbols states emission alphabet using the
> setDistribution method. This Distribution will be your predetermined
> weights.
> >
> > To set the transition probabilities you can use the setWeights(State
> source, Distribution weights). The source is the state you are
> > transitioning from and the weights is the probability of transitioning
> to any State that the source connects too. Because States implement
> Symbol you can put them in a Distribution.
> >
> > To make a Distribution of States that state 'a' could connect to use the
> following pseudo code:
> >
> > State a;
> > Model m;
> > FiniteAlphabet endPoints;
> >
> > endPoints = m.transitionsFrom(a);
> > Distribution d =
> > DistributionFactory.DEFAULT.createDistribution(endPoints);
> >
> > //You can then train d or set it's weights and put it back in the model
> with
> >
> > m.setWeights(a, d);
> >
> > Mark Schreiber
> > Principal Scientist (Bioinformatics)
> >
> > Novartis Institute for Tropical Diseases (NITD)
> > 1 Science Park Road
> > #04-14 The Capricorn, Science Park II
> > Singapore 117528
> >
> > phone +65 6722 2973
> > fax  +65 6722 2910
> >
> >
> >
> >
> >
> > sacoca@mcb.mcgill.ca
> > Sent by: biojava-l-bounces@portal.open-bio.org
> > 03/12/2004 06:11 AM
> >
> >
> >         To:     "Biojava Mailing List" <biojava-l@biojava.org>
> >         cc:
> >         Subject:        [Biojava-l] Parameter Settings in
> > BaumWelchTraining
> >
> >
> > Hi all. I'm trying to optimize the transition states probabilities for
> my HMM. I already have set them to values which I think are pretty good.
> Since I know the Baum Welch can only help with the scores and optimize
> them up to a local maxima I thought of using the parameters I calculated
> as a starting point. The problem is that I don't know how!
> > I followed the example in biojava:
> >
> > ....
> > //train the model to have uniform parameters
> >     ModelTrainer mt = new SimpleModelTrainer();
> >     //register the model to train
> >     mt.registerModel(hmm);
> >
> > I want to use the values already set in my hmm  as the starting
> parameters in the BaumWelch.  I don't want to use the uniform
> distribution as indicated below!
> >
> >     //as no other counts are being used the null weight will cause
> > everything to be uniform
> >     mt.setNullModelWeight(1.0);
> >     mt.train();
> >
> > I tried adding counts and looking up examples on the net but ended up
> more confused than I started. How do I use the addCounts to make this
> work!
> >
> > Stephane Acoca
> > Master's Student
> > McGill Center for Bioinformatics
> >
> > _______________________________________________
> > Biojava-l mailing list  -  Biojava-l@biojava.org
> > http://biojava.org/mailman/listinfo/biojava-l
> >
> >
> >
> 
> 
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 
> 
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 

From mark.schreiber at group.novartis.com  Fri Mar 12 03:58:23 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Fri Mar 12 04:04:04 2004
Subject: [Biojava-l] Parameter Settings in BaumWelchTraining]
Message-ID: <OF6C9DB1F6.2C544EA4-ON48256E55.0031058C-48256E55.00314A88@EU.novartis.net>

I agree. 

If the BaumWelch trainer does cause problems one could always implement a 
different version of ModelTrainer.

- Mark





Dan Bolser <dmb@mrc-dunn.cam.ac.uk>
03/12/2004 04:47 PM

 
        To:     Mark Schreiber/GP/Novartis@PH
        cc:     sacoca@mcb.mcgill.ca, Biojava Mailing List <biojava-l@biojava.org>
        Subject:        Re: [Biojava-l] Parameter Settings in BaumWelchTraining]


On Fri, 12 Mar 2004 mark.schreiber@group.novartis.com wrote:

> When you call the train() method of the BaumWelchTrainer you supply it 
> with a SequenceDB. The sequences from this DB are used to optimize the 
> weights of the model.
> 
> However, I have a bad feeling that when you train your model with the 
> BaumWelchTrainer your previously set counts will be ignored and 
> overwritten. You could check by looking into 
AbstractModelTrainer.train() 
> (which is what the BaumWelchTrainer extends). You could also run some 
> tests to see if using a pre-trained model makes any difference to the 
> final outcome. Does anyone more expert than me on the DP package (ie 
most 
> people) know if the counts are overwritten?

The idea sounds good either way, so it would be a shame to have to reject
it on the basis of a technicality :)

Cheers

> 
> - Mark
> 
> 
> 
> 
> 
> sacoca@mcb.mcgill.ca
> Sent by: biojava-l-bounces@portal.open-bio.org
> 03/12/2004 01:30 PM
> 
> 
>         To:     sacoca@mcb.mcgill.ca
>         cc:     Biojava Mailing List <biojava-l@biojava.org>
>         Subject:        Re: [Biojava-l] Parameter Settings in 
BaumWelchTraining]
> 
> 
> Sorry for the previous error.
> ---------------------------- Original Message 
----------------------------
> Subject: Re: [Biojava-l] Parameter Settings in BaumWelchTraining
> From:    sacoca@MCB.McGill.CA
> Date:    Fri, March 12, 2004 12:27 am
> To:      mark.schreiber@group.novartis.com
> 
--------------------------------------------------------------------------
> 
> Here is the code I have for the training. Using what you told me below, 
I
> can retreive all of the weights that I calculated manually for the hmm
> (distributions for the transitions and distributions for the alphabet of 

> each state). What I do not understand is how to use this information and
> the sequences stored in a file to run the BaumWelchAlgorithm and then
> retreive the optimized values calculated by the algorithm to set them 
back
> into my HMM.
> 
> //Retreive the alphabet of all states
> FiniteAlphabet SA = hmm.stateAlphabet();
> Iterator i = SA.iterator();
> 
> SimpleModelTrainer MT = new SimpleModelTrainer();
> MT.registerModel(hmm);
> 
> //go through each state
> while(i.hasNext())
> {Symbol Currentstate = (Symbol)i.next();
> 
>  //Retreive the distribution of all transitions from the current state
> FiniteAlphabet From = hmm.transitionsFrom((State)Currentstate);
>  Distribution d = hmm.getWeights((State)Currentstate);
>  Iterator i2 = From.iterator();
> 
>  //go through it and look at all the weights for each of the transitions
> while(i2.hasNext())
>     {Symbol s = (Symbol)i2.next();
>      System.out.println("From state "+Currentstate.getName()+
>                         "To State   "+s.getName()+
>                          "Weight     "+d.getWeight(s));}
> 
>  //get the distribution for the alphabet of the current state
>  Distribution d2 =((EmissionState)Currentstate).getDistribution();
> FiniteAlphabet IN = (FiniteAlphabet)hmm.emissionAlphabet();
>  Iterator i3 = IN.iterator();
>  //you can go through it the same way as above using a while loop
> *****************************************************************
> This is what I don't understand!!!!
> *****************************************************************
> here, we have a set of training sequences stored in a file in fasta 
format
> that i'd like to use with the BaumWelch algorithm to optimize the
> transition distributions mentionned above.
> 
> //This is the file with all the training sequences
> BufferedInputStream is = new BufferedInputStream(new
> FileInputStream("z:/Sequences.faa"));
> 
> //Load the file with the SequenceDB class
> SequenceDB DB = SeqIOTools.readFasta(is, ProtAlphabet);
> 
> //use 100 cycles as the stop criteria
> StoppingCriteria stopper = new StoppingCriteria()
>      {public boolean isTrainingComplete(TrainingAlgorithm ta)
>        {return (ta.getCycle() > 100);}};
> 
> *****************************************
> This part is what I am clueless about
> *****************************************
> //How do I optimize my hmm with the BaumWelch algorithm and retreive 
//the
> optimized values ? How do I train the distribution above with //the baum
> welch and the sequences that I have ?
> DP dp= DPFactory.DEFAULT.createDP(hmm);
> BaumWelchTrainer bwt = new BaumWelchTrainer(dp);
> }
> 
> PS : I do not know why you are helping all of us here but thank you. It
> makes Biojava a lot easier to deal with.
> 
> Steve
> 
> > Hi Stephane -
> >
> > Within EmissionState you can set a Distribution that contains emission
> probabilities for the Symbols states emission alphabet using the
> setDistribution method. This Distribution will be your predetermined
> weights.
> >
> > To set the transition probabilities you can use the setWeights(State
> source, Distribution weights). The source is the state you are
> > transitioning from and the weights is the probability of transitioning
> to any State that the source connects too. Because States implement
> Symbol you can put them in a Distribution.
> >
> > To make a Distribution of States that state 'a' could connect to use 
the
> following pseudo code:
> >
> > State a;
> > Model m;
> > FiniteAlphabet endPoints;
> >
> > endPoints = m.transitionsFrom(a);
> > Distribution d =
> > DistributionFactory.DEFAULT.createDistribution(endPoints);
> >
> > //You can then train d or set it's weights and put it back in the 
model
> with
> >
> > m.setWeights(a, d);
> >
> > Mark Schreiber
> > Principal Scientist (Bioinformatics)
> >
> > Novartis Institute for Tropical Diseases (NITD)
> > 1 Science Park Road
> > #04-14 The Capricorn, Science Park II
> > Singapore 117528
> >
> > phone +65 6722 2973
> > fax  +65 6722 2910
> >
> >
> >
> >
> >
> > sacoca@mcb.mcgill.ca
> > Sent by: biojava-l-bounces@portal.open-bio.org
> > 03/12/2004 06:11 AM
> >
> >
> >         To:     "Biojava Mailing List" <biojava-l@biojava.org>
> >         cc:
> >         Subject:        [Biojava-l] Parameter Settings in
> > BaumWelchTraining
> >
> >
> > Hi all. I'm trying to optimize the transition states probabilities for
> my HMM. I already have set them to values which I think are pretty good.
> Since I know the Baum Welch can only help with the scores and optimize
> them up to a local maxima I thought of using the parameters I calculated
> as a starting point. The problem is that I don't know how!
> > I followed the example in biojava:
> >
> > ....
> > //train the model to have uniform parameters
> >     ModelTrainer mt = new SimpleModelTrainer();
> >     //register the model to train
> >     mt.registerModel(hmm);
> >
> > I want to use the values already set in my hmm  as the starting
> parameters in the BaumWelch.  I don't want to use the uniform
> distribution as indicated below!
> >
> >     //as no other counts are being used the null weight will cause
> > everything to be uniform
> >     mt.setNullModelWeight(1.0);
> >     mt.train();
> >
> > I tried adding counts and looking up examples on the net but ended up
> more confused than I started. How do I use the addCounts to make this
> work!
> >
> > Stephane Acoca
> > Master's Student
> > McGill Center for Bioinformatics
> >
> > _______________________________________________
> > Biojava-l mailing list  -  Biojava-l@biojava.org
> > http://biojava.org/mailman/listinfo/biojava-l
> >
> >
> >
> 
> 
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 
> 
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 




From td2 at sanger.ac.uk  Fri Mar 12 04:20:40 2004
From: td2 at sanger.ac.uk (Thomas Down)
Date: Fri Mar 12 04:26:16 2004
Subject: [Biojava-l] Parameter Settings in BaumWelchTraining]
In-Reply-To: <OF092DE640.8908A512-ON48256E55.00205201-48256E55.0021064A@EU.novartis.net>
References: <OF092DE640.8908A512-ON48256E55.00205201-48256E55.0021064A@EU.novartis.net>
Message-ID: <87D428DA-7406-11D8-BF8A-000A95C8B056@sanger.ac.uk>


On 12 Mar 2004, at 06:00, mark.schreiber@group.novartis.com wrote:

> When you call the train() method of the BaumWelchTrainer you supply it
> with a SequenceDB. The sequences from this DB are used to optimize the
> weights of the model.
>
> However, I have a bad feeling that when you train your model with the
> BaumWelchTrainer your previously set counts will be ignored and
> overwritten. You could check by looking into 
> AbstractModelTrainer.train()
> (which is what the BaumWelchTrainer extends). You could also run some
> tests to see if using a pre-trained model makes any difference to the
> final outcome. Does anyone more expert than me on the DP package (ie 
> most
> people) know if the counts are overwritten?

The Baum-Welch algorithm is actually the Expectation Maximization 
algorithm applied to HMMs.  When you run BaumWelchTrainer, it takes the 
existing model then calculates a matrix which defines the probability 
that each symbol in your training database was emitted by a particular 
state in the HMM.  It then optimizes the transition and emission 
probabilities of the model to maximize the likelihood of the data 
*given that assignment of states to data*.

This means that, although all the model parameters get overwritten 
every cycle, they do still depend on the previous state of the model.  
If you're starting with a model that is a good, but not quite optimal, 
fit to your data, BaumWelchTrainer ought to do what you want.

Another thing you might look at is building a model with a mixture of 
trainable and untrainable distributions.  This allows you to 
specifically optimize the bits of the model which you aren't sure about 
yet, while holding the bits which you already trust constant.

      Thomas.

From matthew_pocock at yahoo.co.uk  Fri Mar 12 06:17:56 2004
From: matthew_pocock at yahoo.co.uk (Matthew Pocock)
Date: Fri Mar 12 06:23:44 2004
Subject: [Biojava-l] BlastXMLParserFacade
In-Reply-To: <43dfea43ae54.43ae5443dfea@lbl.gov>
References: <43dfea43ae54.43ae5443dfea@lbl.gov>
Message-ID: <40519C64.5030700@yahoo.co.uk>

Hi,

The parser shouldn't be throwing a NPE. However, the NCBI blast xml 
output didn't used to be well-formed XML, so was not parseable by any 
XML parser. I don't know if this has since been fixed by the NCBI.

Matthew

DMGoodstein@lbl.gov wrote:

>I was wondering if anyone has successfully gotten
>BlastXMLParserFacade to work on an xml style NCBI
>blast output file (version 2.2.3)? I'm getting null
>pointer exceptions deep within the crimson parser
>implementation classes, but I now i'm correctly
>passing in the NCBI output file, since the dtd is
>getting read.
>
>--David Goodstein
>  Joint Genome Institute
>
>/usr/java/j2sdk1.4.1_02/bin/java2/bin/java
>-classpath .:./biojava-1.3.1
>.1.jar BlastParser fred.xml
>a/j2sdk1.4.1_02/bin/javac -classpath .:./biojava-1.3.
>java.lang.NullPointerException
>	at
>org.apache.crimson.parser.Parser2.parseInternal(Parser2.java:524)
>	at
>org.apache.crimson.parser.Parser2.parse(Parser2.java:305)
>	at
>org.apache.crimson.parser.XMLReaderImpl.parse(XMLReaderImpl.java:442)
>	at
>org.biojava.bio.program.sax.blastxml.BlastXMLParserFacade.parse(BlastXMLParserFacade.java:172)
>	at BlastParser.main(BlastParser.java:44)
>1.jar BlastParser
>./fred.xmlava/j2sdk1.4.1_02/bin/java -classpath
>.:./biojava-1.3.1
>java.lang.NullPointerException
>	at
>org.apache.crimson.parser.Parser2.parseInternal(Parser2.java:524)
>	at
>org.apache.crimson.parser.Parser2.parse(Parser2.java:305)
>	at
>org.apache.crimson.parser.XMLReaderImpl.parse(XMLReaderImpl.java:442)
>	at
>org.biojava.bio.program.sax.blastxml.BlastXMLParserFacade.parse(BlastXMLParserFacade.java:172)
>	at BlastParser.main(BlastParser.java:44)
>
>_______________________________________________
>Biojava-l mailing list  -  Biojava-l@biojava.org
>http://biojava.org/mailman/listinfo/biojava-l
>
>  
>

From smh1008 at cus.cam.ac.uk  Fri Mar 12 07:03:37 2004
From: smh1008 at cus.cam.ac.uk (David Huen)
Date: Fri Mar 12 07:09:30 2004
Subject: [Biojava-l] BlastXMLParserFacade
In-Reply-To: <40519C64.5030700@yahoo.co.uk>
References: <43dfea43ae54.43ae5443dfea@lbl.gov> <40519C64.5030700@yahoo.co.uk>
Message-ID: <200403121203.37225.smh1008@cus.cam.ac.uk>

On Friday 12 Mar 2004 11:17 am, Matthew Pocock wrote:
> Hi,
>
> The parser shouldn't be throwing a NPE. However, the NCBI blast xml
> output didn't used to be well-formed XML, so was not parseable by any
> XML parser. I don't know if this has since been fixed by the NCBI.
>
In this case, that's not the problem.  The problem started when we switched 
from the Xerces parsers to Sun's own and no attempt on my part was 
successful in getting it it use the NCBI's DTD.  Part of the problem is 
that DTD from NCBI refers to other DTDs and those references are not 
complete so my code used to explicitly resolve them but with Sun's parser, 
I could never convince it to use the provided DTDs.

I don't have time to return to that problem for a while so perhaps someone 
else more familiar with Crimson could figure it out.  Alternatively, we 
could hack it to not use DTDs at all (totally non-validating).

Regards,
David

From dmb at mrc-dunn.cam.ac.uk  Fri Mar 12 07:19:43 2004
From: dmb at mrc-dunn.cam.ac.uk (Dan Bolser)
Date: Fri Mar 12 07:22:34 2004
Subject: [Biojava-l] BlastXMLParserFacade
In-Reply-To: <40519C64.5030700@yahoo.co.uk>
Message-ID: <Pine.LNX.4.21.0403121216420.5204-100000@mail.mrc-dunn.cam.ac.uk>

On Fri, 12 Mar 2004, Matthew Pocock wrote:

> Hi,
> 
> The parser shouldn't be throwing a NPE. However, the NCBI blast xml 
> output didn't used to be well-formed XML, so was not parseable by any 
> XML parser. I don't know if this has since been fixed by the NCBI.

! 


I have a perl module to 'clean up' large amounts of psi-blast data. Do
such modules exist in biojava?

Specifically, given a 'family' label for query database, I wanted a
'family' assignment to the target database, and my module did this by
assessing overlap within and between families. 

It would be really cool to do this a bit more systematically and
'modularly' with biojava.

 


> 
> Matthew
> 
> DMGoodstein@lbl.gov wrote:
> 
> >I was wondering if anyone has successfully gotten
> >BlastXMLParserFacade to work on an xml style NCBI
> >blast output file (version 2.2.3)? I'm getting null
> >pointer exceptions deep within the crimson parser
> >implementation classes, but I now i'm correctly
> >passing in the NCBI output file, since the dtd is
> >getting read.
> >
> >--David Goodstein
> >  Joint Genome Institute
> >
> >/usr/java/j2sdk1.4.1_02/bin/java2/bin/java
> >-classpath .:./biojava-1.3.1
> >.1.jar BlastParser fred.xml
> >a/j2sdk1.4.1_02/bin/javac -classpath .:./biojava-1.3.
> >java.lang.NullPointerException
> >	at
> >org.apache.crimson.parser.Parser2.parseInternal(Parser2.java:524)
> >	at
> >org.apache.crimson.parser.Parser2.parse(Parser2.java:305)
> >	at
> >org.apache.crimson.parser.XMLReaderImpl.parse(XMLReaderImpl.java:442)
> >	at
> >org.biojava.bio.program.sax.blastxml.BlastXMLParserFacade.parse(BlastXMLParserFacade.java:172)
> >	at BlastParser.main(BlastParser.java:44)
> >1.jar BlastParser
> >./fred.xmlava/j2sdk1.4.1_02/bin/java -classpath
> >.:./biojava-1.3.1
> >java.lang.NullPointerException
> >	at
> >org.apache.crimson.parser.Parser2.parseInternal(Parser2.java:524)
> >	at
> >org.apache.crimson.parser.Parser2.parse(Parser2.java:305)
> >	at
> >org.apache.crimson.parser.XMLReaderImpl.parse(XMLReaderImpl.java:442)
> >	at
> >org.biojava.bio.program.sax.blastxml.BlastXMLParserFacade.parse(BlastXMLParserFacade.java:172)
> >	at BlastParser.main(BlastParser.java:44)
> >
> >_______________________________________________
> >Biojava-l mailing list  -  Biojava-l@biojava.org
> >http://biojava.org/mailman/listinfo/biojava-l
> >
> >  
> >
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
> 

From td2 at sanger.ac.uk  Fri Mar 12 07:27:02 2004
From: td2 at sanger.ac.uk (Thomas Down)
Date: Fri Mar 12 07:32:38 2004
Subject: [Biojava-l] BlastXMLParserFacade
In-Reply-To: <200403121203.37225.smh1008@cus.cam.ac.uk>
References: <43dfea43ae54.43ae5443dfea@lbl.gov> <40519C64.5030700@yahoo.co.uk>
	<200403121203.37225.smh1008@cus.cam.ac.uk>
Message-ID: <911015F5-7420-11D8-BF8A-000A95C8B056@sanger.ac.uk>


On 12 Mar 2004, at 12:03, David Huen wrote:

> On Friday 12 Mar 2004 11:17 am, Matthew Pocock wrote:
>> Hi,
>>
>> The parser shouldn't be throwing a NPE. However, the NCBI blast xml
>> output didn't used to be well-formed XML, so was not parseable by any
>> XML parser. I don't know if this has since been fixed by the NCBI.
>>
> In this case, that's not the problem.  The problem started when we 
> switched
> from the Xerces parsers to Sun's own and no attempt on my part was
> successful in getting it it use the NCBI's DTD.  Part of the problem is
> that DTD from NCBI refers to other DTDs and those references are not
> complete so my code used to explicitly resolve them but with Sun's 
> parser,
> I could never convince it to use the provided DTDs.
>
> I don't have time to return to that problem for a while so perhaps 
> someone
> else more familiar with Crimson could figure it out.  Alternatively, we
> could hack it to not use DTDs at all (totally non-validating).

We're not really using Sun's parser specifically -- just the default 
JAXP parser that's installed on the system.  It ought to be possible to 
make Xerces the latest JAXP parser, although I think that this means 
putting it on the bootstrap classpath.  I'll try this out.

A while ago I was under the distinct impression that Sun were going to 
junk Crimson and make Xerces the standard parser.  Looks like this 
hasn't happened though -- I don't know why.

      Thomas.

From fangl at genomics.org.cn  Fri Mar 19 01:11:12 2004
From: fangl at genomics.org.cn (Magic Fang)
Date: Fri Mar 19 01:20:22 2004
Subject: [Biojava-l] is there a biojava class diagram? like bioperl diagram
Message-ID: <405A8F00.40602@genomics.org.cn>


From fangl at genomics.org.cn  Fri Mar 19 01:37:51 2004
From: fangl at genomics.org.cn (Magic Fang)
Date: Fri Mar 19 01:47:19 2004
Subject: [Biojava-l] yes, something like this
Message-ID: <405A953F.4020807@genomics.org.cn>

this one seems too detail and complex, i mean just class function and 
inherit relationship.
From mark.schreiber at group.novartis.com  Fri Mar 19 03:26:08 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Fri Mar 19 03:32:16 2004
Subject: [Biojava-l] yes, something like this
Message-ID: <OFCE22BD4B.51E3CD4C-ON48256E5C.002DC360-48256E5C.002E56E8@EU.novartis.net>

Unfortunately the heavy use of interfaces and 'mutliple inheritance' in 
biojava means that those UML diagrams are about as simple as you will get. 
I suppose they could be pruned a bit to show only interfaces implemted and 
classes extended and not dependencies and reverse dependencies.

- Mark





Magic Fang <fangl@genomics.org.cn>
Sent by: biojava-l-bounces@portal.open-bio.org
03/19/2004 02:37 PM

 
        To:     Mark Schreiber/GP/Novartis@PH
        cc:     biojava-l@biojava.org
        Subject:        [Biojava-l] yes, something like this


this one seems too detail and complex, i mean just class function and 
inherit relationship.
_______________________________________________
Biojava-l mailing list  -  Biojava-l@biojava.org
http://biojava.org/mailman/listinfo/biojava-l



From matthew_pocock at yahoo.co.uk  Fri Mar 19 06:12:39 2004
From: matthew_pocock at yahoo.co.uk (Matthew Pocock)
Date: Fri Mar 19 06:18:21 2004
Subject: [Biojava-l] yes, something like this
In-Reply-To: <OFCE22BD4B.51E3CD4C-ON48256E5C.002DC360-48256E5C.002E56E8@EU.novartis.net>
References: <OFCE22BD4B.51E3CD4C-ON48256E5C.002DC360-48256E5C.002E56E8@EU.novartis.net>
Message-ID: <405AD5A7.9080606@yahoo.co.uk>

Agreed - biojava and UML don't like each other. UML seems much more 
suited to data-blob designs where as biojava is mainly a relationships 
and associations design. The OWL for biojava looks ok though. Hey ho.

mark.schreiber@group.novartis.com wrote:

>Unfortunately the heavy use of interfaces and 'mutliple inheritance' in 
>biojava means that those UML diagrams are about as simple as you will get. 
>I suppose they could be pruned a bit to show only interfaces implemted and 
>classes extended and not dependencies and reverse dependencies.
>
>- Mark
>
>
>
>
>
>Magic Fang <fangl@genomics.org.cn>
>Sent by: biojava-l-bounces@portal.open-bio.org
>03/19/2004 02:37 PM
>
> 
>        To:     Mark Schreiber/GP/Novartis@PH
>        cc:     biojava-l@biojava.org
>        Subject:        [Biojava-l] yes, something like this
>
>
>this one seems too detail and complex, i mean just class function and 
>inherit relationship.
>_______________________________________________
>Biojava-l mailing list  -  Biojava-l@biojava.org
>http://biojava.org/mailman/listinfo/biojava-l
>
>
>
>_______________________________________________
>Biojava-l mailing list  -  Biojava-l@biojava.org
>http://biojava.org/mailman/listinfo/biojava-l
>
>  
>

From heuermh at acm.org  Fri Mar 19 12:28:58 2004
From: heuermh at acm.org (Michael Heuer)
Date: Fri Mar 19 12:43:32 2004
Subject: [Biojava-l] yes, something like this
In-Reply-To: <405AD5A7.9080606@yahoo.co.uk>
Message-ID: <Pine.GSO.4.44.0403191227430.20572-100000@shell3.shore.net>


On Fri, 19 Mar 2004, Matthew Pocock wrote:

> The OWL for biojava looks ok though. Hey ho.

Is this a tease? :)  What would/could one do with an OWL representation of
the biojava object model?

   michael

From fangl at genomics.org.cn  Fri Mar 19 23:57:09 2004
From: fangl at genomics.org.cn (Magic Fang)
Date: Sat Mar 20 00:02:29 2004
Subject: [Biojava-l] yes, something like this
In-Reply-To: <Pine.GSO.4.44.0403191227430.20572-100000@shell3.shore.net>
References: <Pine.GSO.4.44.0403191227430.20572-100000@shell3.shore.net>
Message-ID: <405BCF25.3030307@genomics.org.cn>

Michael Heuer wrote:

>On Fri, 19 Mar 2004, Matthew Pocock wrote:
>
>  
>
>>The OWL for biojava looks ok though. Hey ho.
>>    
>>
>
>Is this a tease? :)  What would/could one do with an OWL representation of
>the biojava object model?
>
>   michael
>
>
>
>  
>
Does the OWL means the Open Wrokflow/Lifecycle model?  I would refer it 
to owl, that few would see it or I am confused on biojava architecture:-)
It is more complex than bioperl.
From heuermh at acm.org  Sat Mar 20 00:35:10 2004
From: heuermh at acm.org (Michael Heuer)
Date: Sat Mar 20 00:44:52 2004
Subject: [Biojava-l] yes, something like this
In-Reply-To: <405BCF25.3030307@genomics.org.cn>
Message-ID: <Pine.GSO.4.44.0403200033540.21294-100000@shell3.shore.net>


On Sat, 20 Mar 2004, Magic Fang wrote:

> Does the OWL means the Open Wrokflow/Lifecycle model?  I would refer it
> to owl, that few would see it or I am confused on biojava architecture:-)
> It is more complex than bioperl.

Nope, this OWL

> http://www.w3.org/TR/owl-features/

   michael

From mark.schreiber at group.novartis.com  Sun Mar 21 21:21:41 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Sun Mar 21 21:27:11 2004
Subject: [Biojava-l] BioJava mentioned in 'on java' article 
Message-ID: <OFDFEE27AE.C126357D-ON48256E5F.000CD135-48256E5F.000CF8FE@EU.novartis.net>

Thanks to Russell Smithies for passing this on:

http://www.onjava.com/pub/a/onjava/2004/03/10/bioinf.html

Mark Schreiber
Principal Scientist (Bioinformatics)

Novartis Institute for Tropical Diseases (NITD)
1 Science Park Road
#04-14 The Capricorn, Science Park II
Singapore 117528

phone +65 6722 2973
fax  +65 6722 2910

From Christian.Gruber at biomax.com  Mon Mar 22 10:30:02 2004
From: Christian.Gruber at biomax.com (Christian Gruber)
Date: Mon Mar 22 10:35:27 2004
Subject: [Biojava-l] Missing ProcessTools class
Message-ID: <405F067A.5040708@biomax.com>

Hi!

I just found that the class org.biojava.utils.ProcessTools is not part 
of the source and binary distribution of BioJava 1.3.1, but they are 
part of the biojava-live CVS. Are they missing by accident?

By the way, what is the recommended way to compile BioJava 1.3.1, since 
biojava-1.3.1.tar.gz is missing build.xml etc. I tried to unpack 1.3.0 
and on top of that the 1.3.1 source, and it seems to work, but I'm not 
sure if this is the correct way to do that...

Greetings,
Christian
From mark.schreiber at group.novartis.com  Mon Mar 22 20:16:43 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Mon Mar 22 20:22:16 2004
Subject: [Biojava-l] Missing ProcessTools class
Message-ID: <OF7B5D7BD9.C879A1D3-ON48256E60.00068F69-48256E60.0007066C@EU.novartis.net>

BioJava 1.3.1 was an update release of 1.3. There are a number of things 
that biojava live contains that 1.3.1 does not. The goal was to release a 
1.3.2 but unfortunately I moved countries and jobs etc so the plan slipped 
somewhat. It now seems unlikely that there will be any further updates of 
the 1.3.x lineage.

If 1.3.1 doesn't have what you need I would reccommend getting one of the 
nightly snapshots of biojava-live (one that passes all the unit tests) and 
build that. The vast majority of the live build is now stable. Only newly 
added things tend to change rapidly.

- Mark





Christian Gruber <Christian.Gruber@biomax.com>
Sent by: biojava-l-bounces@portal.open-bio.org
03/22/2004 11:30 PM

 
        To:     biojava-l@biojava.org
        cc: 
        Subject:        [Biojava-l] Missing ProcessTools class


Hi!

I just found that the class org.biojava.utils.ProcessTools is not part 
of the source and binary distribution of BioJava 1.3.1, but they are 
part of the biojava-live CVS. Are they missing by accident?

By the way, what is the recommended way to compile BioJava 1.3.1, since 
biojava-1.3.1.tar.gz is missing build.xml etc. I tried to unpack 1.3.0 
and on top of that the 1.3.1 source, and it seems to work, but I'm not 
sure if this is the correct way to do that...

Greetings,
Christian
_______________________________________________
Biojava-l mailing list  -  Biojava-l@biojava.org
http://biojava.org/mailman/listinfo/biojava-l



From fangl at genomics.org.cn  Mon Mar 22 20:22:38 2004
From: fangl at genomics.org.cn (Magic Fang)
Date: Mon Mar 22 20:32:35 2004
Subject: [Biojava-l] is blast parser flawed?
Message-ID: <405F915E.8090809@genomics.org.cn>

it seems that blast parser only check blast version on the first line of 
blast output, and i tricked her by changing the first line to 2.2.3, 
while i use version of 2.2.8.
From mark.schreiber at group.novartis.com  Mon Mar 22 20:53:30 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Mon Mar 22 20:59:23 2004
Subject: [Biojava-l] is blast parser flawed?
Message-ID: <OFF0BA6281.B9B7F308-ON48256E60.000A1CF0-48256E60.000A64A8@EU.novartis.net>

Hi -

The BlastLikeSAXParser checks the version of blast against a list it is 
known to support. If you call the .setModeLazy() method then the parser 
doesn't check the version. If you do this you should check the results 
carefully.

If people know of BLAST versions that the parser reads reliably could they 
let us know and we can add them to the list of supported versions.

- Mark





Magic Fang <fangl@genomics.org.cn>
Sent by: biojava-l-bounces@portal.open-bio.org
03/23/2004 09:22 AM

 
        To:     biojava-l@biojava.org
        cc: 
        Subject:        [Biojava-l] is blast parser flawed?


it seems that blast parser only check blast version on the first line of 
blast output, and i tricked her by changing the first line to 2.2.3, 
while i use version of 2.2.8.
_______________________________________________
Biojava-l mailing list  -  Biojava-l@biojava.org
http://biojava.org/mailman/listinfo/biojava-l



From Christian.Gruber at biomax.com  Tue Mar 23 09:02:18 2004
From: Christian.Gruber at biomax.com (Christian Gruber)
Date: Tue Mar 23 09:07:42 2004
Subject: [Biojava-l] cvs.biojava.org down?
Message-ID: <4060436A.4090902@biomax.com>

Hi!

I can't access the host cvs.biojava.org via ping and http. Maybe the 
machine is down?

Christian
From max_dipl at web.de  Thu Mar 25 10:22:02 2004
From: max_dipl at web.de (Maximilian Haeussler)
Date: Thu Mar 25 10:27:15 2004
Subject: [Biojava-l] LabelledSequenceRenderer does not work with
 TranslatedSequenceRenderer, Newbie
Message-ID: <4062F91A.8020306@web.de>

Hallo everyone,

I am pretty new to biojava. Can anyone give me a hint why the 
LabelledSequenceRenderer works fine for the SequenceRenderer, but not 
with the TranslatedSequenceRenderer? I've modified FastBeadDemo and 
BeadDemo to include labels. BeadDemo works fine but not the translated 
one. What is so different about the TranslatedSequenceRenderer that 
prohibits painting into the trailing space of the sequence?

Another question, this time a rather stupid one: How can I get rid of 
those "Originally:"/"Now:"-messages of the demos? I can't find the 
message anywhere, neither in biojava nor in the demo directory...

Thanks
Max

From dmb at mrc-dunn.cam.ac.uk  Thu Mar 25 10:49:09 2004
From: dmb at mrc-dunn.cam.ac.uk (Dan Bolser)
Date: Thu Mar 25 10:49:05 2004
Subject: [Biojava-l] biojava script?
Message-ID: <Pine.LNX.4.21.0403251547120.31968-100000@mail.mrc-dunn.cam.ac.uk>


Hello,

I am hoping to write a script to parse HMMER results and display the
results. I know modules to do both these things exist, but has anyone put
them together in a script yet?

Cheers,
Dan.

From kalle.naslund at genpat.uu.se  Thu Mar 25 11:25:43 2004
From: kalle.naslund at genpat.uu.se (=?ISO-8859-1?Q?Kalle_N=E4slund?=)
Date: Thu Mar 25 11:31:20 2004
Subject: [Biojava-l] LabelledSequenceRenderer does not work with
	TranslatedSequenceRenderer, Newbie
In-Reply-To: <4062F91A.8020306@web.de>
References: <4062F91A.8020306@web.de>
Message-ID: <40630807.40000@genpat.uu.se>

Maximilian Haeussler wrote:

> Hallo everyone,
>
> I am pretty new to biojava. Can anyone give me a hint why the 
> LabelledSequenceRenderer works fine for the SequenceRenderer, but not 
> with the TranslatedSequenceRenderer? I've modified FastBeadDemo and 
> BeadDemo to include labels. BeadDemo works fine but not the translated 
> one. What is so different about the TranslatedSequenceRenderer that 
> prohibits painting into the trailing space of the sequence?
>
> Another question, this time a rather stupid one: How can I get rid of 
> those "Originally:"/"Now:"-messages of the demos? I can't find the 
> message anywhere, neither in biojava nor in the demo directory...
>
> Thanks
> Max
>
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l


HI!

Firstly, i am going to make a  few assumptions here, as some of the 
classnames you mention seems a bit out of context
or do not exist. My appologies if these assumptions are wrong, i will 
then have to blame my lack of proper english
knowledge. But, hope i assumed correctly and this helps.

The assumption are.

1) SequenceRenderer                   == SequencePanel
   ( if we are going to compare this class to TranslatedSequencePanel, 
SequencePanel is the thing that makes sense )
  
2) TranslatedSequenceRenderer         == TranslatedSequencePanel
   ( there is no TranslatedSequenceRenderer in the bj source tree )

3) The part about painting in the trailing space of the sequence is 
realy about the leading space   
   ( The LabelledSequenceRenderer doesnt support painting into the 
trailing space, as indicated by the javadocs )
     This should of course be fixed )



then then the question becomes, why does LabelledSequenceRenderer work 
with SequencePanel and not with TranslatedSequencePanel ?

My guess is that this is a bug in either the SequencePanel or the 
TranslatedSequencePanel when it comes to handling
leading and trailing space.

The obvious solutions as i see it is

1) you use SequencePanel, after a quick look at the SequencePanel api it 
seems you can use the setRange() method call
   to mimic the TranslatedSequencePanel setSymbolTranslation(int 
translation)  behaviour.
   by something like setRange(  new RangeLocation( translation, 
symbolList.length() ) );

2) have me or someone else fix 
TranslatedSEquencePanel/SequencePanel/LabelledSequenceRenderer properly
   ( this should realy be done irrespectively of what solution you choose )

I will be able to assist you with point 2, but not until sometime in the 
beginning of next week.

Also, can anyone remind me why we have several Panel implementations 
please ? shouldnt it be very nice to have one single
implementation, that supports the unique features of both SequencePanel 
and TranslatedSequencePanel ?

Live long and prosper
/Kalle



From matthew_pocock at yahoo.co.uk  Thu Mar 25 11:30:53 2004
From: matthew_pocock at yahoo.co.uk (Matthew Pocock)
Date: Thu Mar 25 11:41:46 2004
Subject: [Biojava-l] LabelledSequenceRenderer does not work with
	TranslatedSequenceRenderer, Newbie
In-Reply-To: <4062F91A.8020306@web.de>
References: <4062F91A.8020306@web.de>
Message-ID: <4063093D.1010709@yahoo.co.uk>

Maximilian Haeussler wrote:

> Hallo everyone,
>

> Another question, this time a rather stupid one: How can I get rid of 
> those "Originally:"/"Now:"-messages of the demos? I can't find the 
> message anywhere, neither in biojava nor in the demo directory...

I found some code in my copy in org.biojava.bio.gui.sequence.GuiTools 
that is printing these messages (lines #33 and #34). These lines are 
commented out in the version in CVS (revision 1.4). Does using code from 
CVS help? If you think you are doing this already, is it possible that 
you have an earlier biojava.jar in your classpath?

Matthew

>
> Thanks
> Max
>
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l
>

From matthew_pocock at yahoo.co.uk  Thu Mar 25 12:20:24 2004
From: matthew_pocock at yahoo.co.uk (Matthew Pocock)
Date: Thu Mar 25 12:26:04 2004
Subject: [Biojava-l] LabelledSequenceRenderer does not work
	with	TranslatedSequenceRenderer, Newbie
In-Reply-To: <40630807.40000@genpat.uu.se>
References: <4062F91A.8020306@web.de> <40630807.40000@genpat.uu.se>
Message-ID: <406314D8.1020805@yahoo.co.uk>

Kalle N?slund wrote:

> Also, can anyone remind me why we have several Panel implementations 
> please ? shouldnt it be very nice to have one single
> implementation, that supports the unique features of both 
> SequencePanel and TranslatedSequencePanel ?

Amen brother! Let's drop the poster class - it's bad. SequencePanel 
behaves itself in a scrollable pane. TranslatedSequencePanel needs to 
have hooks added to scroll it. Is there any performance difference 
between the two now? Does one render things better than the other e.g. 
on long sequences? Let's drop one of these and go with the other.

Kalle - want to proposing what to do next?

>
> Live long and prosper
> /Kalle


Matthew
From kdj at sanger.ac.uk  Thu Mar 25 12:26:46 2004
From: kdj at sanger.ac.uk (Keith James)
Date: Thu Mar 25 12:32:05 2004
Subject: [Biojava-l] LabelledSequenceRenderer does not work with
	TranslatedSequenceRenderer, Newbie
In-Reply-To: <40630807.40000@genpat.uu.se>
References: <4062F91A.8020306@web.de> <40630807.40000@genpat.uu.se>
Message-ID: <sc48yhozvl5.fsf@hgs2b.internal.sanger.ac.uk>

>>>>> "Kalle" == Kalle N?slund <kalle.naslund@genpat.uu.se> writes:

[...]

    Kalle> Also, can anyone remind me why we have several Panel
    Kalle> implementations please ? shouldnt it be very nice to have
    Kalle> one single implementation, that supports the unique
    Kalle> features of both SequencePanel and TranslatedSequencePanel
    Kalle> ?

Long story...

This came about because SequencePanel had two problems:

1. It will by default render a huge sequence in a single huge
   coordinate space which leads to floating point errors in marking
   tics/points at the far end of the sequence.

   To solve this a hack was put in to translate the drawing
   coordinates to ~0 before drawing. However that broke rendering to
   any Graphics2D derived from a BufferedImage where it produced an
   un-renderable region between 0 and the offset (which was 50 pixels
   wide, if I remember)

2. It was far too slow with available sequence/feature renderers.


So I wrote TranslatedSequencePanel which solves the first problem
without using the offset hack. Instead of placing a huge virtual
rendering area in a ScrollPane, it clips a window of sequence to
render and runs the window along the sequence. Therefore you are
always drawing from 0, even at base 100,000,001 of the sequence.

As a nice side-effect, it also turned out to be 5-10x faster at
rendering with the current sequence/feature renderers.

But people were still using the SequencePanel in their client code, so
I didn't want to remove it.

Keith

-- 

- Keith James <kdj@sanger.ac.uk> Microarray Facility, Team 65 -
- The Wellcome Trust Sanger Institute, Hinxton, Cambridge, UK -

From smh1008 at cus.cam.ac.uk  Thu Mar 25 12:45:10 2004
From: smh1008 at cus.cam.ac.uk (David Huen)
Date: Thu Mar 25 12:59:32 2004
Subject: [Biojava-l] LabelledSequenceRenderer does not work with
	TranslatedSequenceRenderer, Newbie
In-Reply-To: <sc48yhozvl5.fsf@hgs2b.internal.sanger.ac.uk>
References: <4062F91A.8020306@web.de> <40630807.40000@genpat.uu.se>
	<sc48yhozvl5.fsf@hgs2b.internal.sanger.ac.uk>
Message-ID: <200403251745.10493.smh1008@cus.cam.ac.uk>

On Thursday 25 Mar 2004 5:26 pm, Keith James wrote:
> >>>>> "Kalle" == Kalle N?slund <kalle.naslund@genpat.uu.se> writes:
>
> [...]
>
>     Kalle> Also, can anyone remind me why we have several Panel
>     Kalle> implementations please ? shouldnt it be very nice to have
>     Kalle> one single implementation, that supports the unique
>     Kalle> features of both SequencePanel and TranslatedSequencePanel
>     Kalle> ?
>
> Long story...
>
> This came about because SequencePanel had two problems:
>
> 1. It will by default render a huge sequence in a single huge
>    coordinate space which leads to floating point errors in marking
>    tics/points at the far end of the sequence.
>
>    To solve this a hack was put in to translate the drawing
>    coordinates to ~0 before drawing. However that broke rendering to
>    any Graphics2D derived from a BufferedImage where it produced an
>    un-renderable region between 0 and the offset (which was 50 pixels
>    wide, if I remember)
>
> 2. It was far too slow with available sequence/feature renderers.
>
>
> So I wrote TranslatedSequencePanel which solves the first problem
> without using the offset hack. Instead of placing a huge virtual
> rendering area in a ScrollPane, it clips a window of sequence to
> render and runs the window along the sequence. Therefore you are
> always drawing from 0, even at base 100,000,001 of the sequence.
>
> As a nice side-effect, it also turned out to be 5-10x faster at
> rendering with the current sequence/feature renderers.
>
> But people were still using the SequencePanel in their client code, so
> I didn't want to remove it.
>
Let's go to TranslatedSequencePanel then.  It seems the sane way to go and 
it is good at high coord spaces.  The only way this is going to happen is 
if we tag it "DEPRECATED" immediately and shoot it at 1.4 final.

David Huen


From mark.schreiber at group.novartis.com  Fri Mar 19 01:25:14 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Thu Mar 25 16:39:23 2004
Subject: [Biojava-l] is there a biojava class diagram? like bioperl diagram
Message-ID: <OFE6C7D635.8F858FE3-ON48256E5C.00232BFF-48256E5C.00234547@EU.novartis.net>

Do you mean UML diagrams?

Attached is an example of the SymbolList UML


-------------- next part --------------
A non-text attachment was scrubbed...
Name: symbol-list-uml.png
Type: application/octet-stream
Size: 11377 bytes
Desc: not available
Url : http://portal.open-bio.org/pipermail/biojava-l/attachments/20040319/b038a990/symbol-list-uml.obj
From fangl at genomics.org.cn  Thu Mar 25 20:26:05 2004
From: fangl at genomics.org.cn (Magic Fang)
Date: Thu Mar 25 20:32:00 2004
Subject: [Biojava-l] what is the design goal of biojava?
Message-ID: <406386AD.5070503@genomics.org.cn>

hi,
i am now using biojava, and i think it is quite good, especially the 
modeling and statistics class, but on data manipulation, it seems 
somewhat complex. if she is like bioperl in seqio or searchio, she will 
be more beautiful i think. so i want the design goal of biojava, for 
biologist or programmer?
From mark.schreiber at group.novartis.com  Fri Mar 26 02:58:33 2004
From: mark.schreiber at group.novartis.com (mark.schreiber@group.novartis.com)
Date: Fri Mar 26 03:04:03 2004
Subject: [Biojava-l] what is the design goal of biojava?
Message-ID: <OF0F5C7308.FA790300-ON48256E63.002B942B-48256E63.002BD05F@EU.novartis.net>

I would say the design goal (not that there is an official one) is general 
sequence manipulation and analysis. BioJava is probably more for the 
developer than the biologist. Part of the reason is the design and part of 
the reason is that Java is more suited to application development than 
simple scripting  which is better done in Perl, Python or Ruby.

- Mark





Magic Fang <fangl@genomics.org.cn>
Sent by: biojava-l-bounces@portal.open-bio.org
03/26/2004 09:26 AM

 
        To:     biojava-l@biojava.org
        cc: 
        Subject:        [Biojava-l] what is the design goal of biojava?


hi,
i am now using biojava, and i think it is quite good, especially the 
modeling and statistics class, but on data manipulation, it seems 
somewhat complex. if she is like bioperl in seqio or searchio, she will 
be more beautiful i think. so i want the design goal of biojava, for 
biologist or programmer?
_______________________________________________
Biojava-l mailing list  -  Biojava-l@biojava.org
http://biojava.org/mailman/listinfo/biojava-l



From kdj at sanger.ac.uk  Fri Mar 26 04:52:51 2004
From: kdj at sanger.ac.uk (Keith James)
Date: Fri Mar 26 04:58:11 2004
Subject: [Biojava-l] LabelledSequenceRenderer does not work
	with	TranslatedSequenceRenderer, Newbie
In-Reply-To: <406314D8.1020805@yahoo.co.uk>
References: <4062F91A.8020306@web.de> <40630807.40000@genpat.uu.se>
	<406314D8.1020805@yahoo.co.uk>
Message-ID: <sc4wu58kk98.fsf@hgs2d.internal.sanger.ac.uk>

>>>>> "Matthew" == Matthew Pocock <matthew_pocock@yahoo.co.uk> writes:

    Matthew> Kalle N?slund wrote:
    >> Also, can anyone remind me why we have several Panel
    >> implementations please ? shouldnt it be very nice to have one
    >> single implementation, that supports the unique features of
    >> both SequencePanel and TranslatedSequencePanel ?

    Matthew> Amen brother! Let's drop the poster class - it's
    Matthew> bad. SequencePanel behaves itself in a scrollable
    Matthew> pane. TranslatedSequencePanel needs to have hooks added
    Matthew> to scroll it. Is there any performance difference between
    Matthew> the two now? Does one render things better than the other
    Matthew> e.g. on long sequences? Let's drop one of these and go
    Matthew> with the other.

That performance difference was on JDK 1.3 (I think) Linux and Tru64.
It could be that with all the changes in 1.4+ that the difference has
closed - one would hope so. I haven't looked at this in a long
while. I will check it out.

Keith

-- 

- Keith James <kdj@sanger.ac.uk> Microarray Facility, Team 65 -
- The Wellcome Trust Sanger Institute, Hinxton, Cambridge, UK -

From matthew_pocock at yahoo.co.uk  Fri Mar 26 08:36:10 2004
From: matthew_pocock at yahoo.co.uk (Matthew Pocock)
Date: Fri Mar 26 08:41:59 2004
Subject: [Biojava-l] what is the design goal of biojava?
In-Reply-To: <406386AD.5070503@genomics.org.cn>
References: <406386AD.5070503@genomics.org.cn>
Message-ID: <406431CA.1030402@yahoo.co.uk>

Magic Fang wrote:

> hi,
> i am now using biojava, and i think it is quite good, especially the 
> modeling and statistics class, but on data manipulation, it seems 
> somewhat complex. if she is like bioperl in seqio or searchio, she 
> will be more beautiful i think. so i want the design goal of biojava, 
> for biologist or programmer?

I agree with you about the complexity. If you can think of a cleaner and 
simpler API and have the time or can motivate others then feel free to 
tackle this. I think this happened because the early developers where 
all using BioJava as a library for developing compute-intensive 
applications, rather than as a scripting environment for running 
external anaysis programs. Obviously, now it would be good to address this.

Matthew

From fangl at genomics.org.cn  Tue Mar 30 01:33:55 2004
From: fangl at genomics.org.cn (Magic Fang)
Date: Tue Mar 30 01:39:33 2004
Subject: [Biojava-l] how to get renderer in SequencePanel
Message-ID: <406914D3.6030407@genomics.org.cn>

hi,
i am now developing a trace viewer and add trace dynamicly, but i want 
to change the AbiTraceRenderer's depth when i need. and how to get 
MultiLineRenderer  and then AbiTaceRenderer?
thank u.
From kalle.naslund at genpat.uu.se  Tue Mar 30 04:05:59 2004
From: kalle.naslund at genpat.uu.se (=?ISO-8859-1?Q?Kalle_N=E4slund?=)
Date: Tue Mar 30 04:10:55 2004
Subject: [Fwd: Re: [Biojava-l] LabelledSequenceRenderer does not work with
	TranslatedSequenceRenderer, Newbie]
Message-ID: <40693877.40107@genpat.uu.se>

Keith James wrote:

> [ ... ]
>
> Long story...
>
> This came about because SequencePanel had two problems:
>
> 1. It will by default render a huge sequence in a single huge
>   coordinate space which leads to floating point errors in marking
>   tics/points at the far end of the sequence.
>
>   To solve this a hack was put in to translate the drawing
>   coordinates to ~0 before drawing. However that broke rendering to
>   any Graphics2D derived from a BufferedImage where it produced an
>   un-renderable region between 0 and the offset (which was 50 pixels
>   wide, if I remember)
>
> 2. It was far too slow with available sequence/feature renderers.
>
>
> So I wrote TranslatedSequencePanel which solves the first problem
> without using the offset hack. Instead of placing a huge virtual
> rendering area in a ScrollPane, it clips a window of sequence to
> render and runs the window along the sequence. Therefore you are
> always drawing from 0, even at base 100,000,001 of the sequence.
>
> As a nice side-effect, it also turned out to be 5-10x faster at
> rendering with the current sequence/feature renderers.
>
> But people were still using the SequencePanel in their client code, so
> I didn't want to remove it.
>
> Keith
>
>  
>
Matthew have done some work on the rendering part, there is now a class
called org.biojava.bio.gui.sequence.GUITools that contains some helper
methods, so that SequenceRenderers can figure out what part is visible, so
they can limit rendering to the visible parts. From a quick test, the speed
with SequencePanel seems comparable to the speed you get with
TranslatedSequencePanel.

I dont know if he has fixed issue number 1 aswell, but i will try to 
take a look
at that later on today, and see if i can see the rendering errors with 
the CVS
version of SequencePanel.

Kalle