From cbonnard at isrec-sg1.unil.ch Fri Nov 1 04:21:20 2002 From: cbonnard at isrec-sg1.unil.ch (Claude Bonnard) Date: Fri, 1 Nov 2002 10:21:20 +0100 Subject: Remote getz from emboss Message-ID: <10211011021.ZM29739@isrec-sg1> Dear, Does someone use a script (shell or perl) to access a srs server on an other machine with rsh? If yes, is it possible to have this script? the idea is to replace the "local" getz by a rgetz. I am aware of the gbwget development, but I am sure that the equivalent for getz has been done somewhere. -Many thanks in advance, regards, Claude -- Claude Bonnard Ph.D. ISREC (Swiss Institute for Experimental Cancer Research) Bioinformatics Group Ch des Boveresses 155 CH-1066 Epalinges Switzerland phone: [41-21]-692-5891 fax: [41-21]-652-6933 From rls at ebi.ac.uk Fri Nov 1 04:47:17 2002 From: rls at ebi.ac.uk (Rodrigo Lopez) Date: Fri, 1 Nov 2002 09:47:17 -0000 Subject: Remote getz from emboss In-Reply-To: <10211011021.ZM29739@isrec-sg1> Message-ID: Hi, (not directly related to rsh...) :-) The EBI's External Services group is working on a web service that uses SOAP and Axis to address this. We have a prototype running in-house that permits us to use the service within EMBOSS as an external method. Sharmila Pilai, who has been working on this can provide more details. From the user point of view things look like this: % seqret wsembl:hscfos\* -auto -out myfile.tfa The above will use the web service to retrieve data from a source such as SRS, server w/ cdrom index files (another emboss installation), etc. The user needs to install a perl or java client on his/her machine an have some bits and bobs to run the SOAP/Axis client part of the service. Another thing we are working on is being able to use EMBOSS remotely using the same technology, for example from a Windows2000 machine using the command prompt. All that would be required is again a perl/java client and general usage looks as follows: C:\My Documents\wsemboss {appname} {params} Obviously, all can be hidden behind batch scripts to make it even simple to use and in effect, design a simple workflow such as: C:\My Documents\wseqret wsembl:hscfos\* -auto -stdout | wnewcpgreport -filter -auto Anyway, I think Sharmila will have some more comments to make about this and tell you more about her progress. Sharmila, I think you should join this list (if you are not already there?). R:) > -----Original Message----- > From: owner-emboss-dev at hgmp.mrc.ac.uk > [mailto:owner-emboss-dev at hgmp.mrc.ac.uk]On Behalf Of Claude Bonnard > Sent: 01 November 2002 09:21 > To: EMBOSS Dev > Subject: Remote getz from emboss > > > Dear, > > Does someone use a script (shell or perl) to access a srs server > on an other > machine with rsh? If yes, is it possible to have this script? > > the idea is to replace the "local" getz by a rgetz. > I am aware of the gbwget development, but I am sure that the > equivalent for > getz has been done somewhere. > -Many thanks in advance, > > regards, > > Claude > > > -- > Claude Bonnard Ph.D. > ISREC (Swiss Institute for Experimental Cancer Research) > Bioinformatics Group > Ch des Boveresses 155 > CH-1066 Epalinges > Switzerland > phone: [41-21]-692-5891 > fax: [41-21]-652-6933 > From cbeazley at hgmp.mrc.ac.uk Tue Nov 12 06:31:01 2002 From: cbeazley at hgmp.mrc.ac.uk (Claude Beazley) Date: Tue, 12 Nov 2002 11:31:01 +0000 Subject: gcc Optimization of code Message-ID: <200211121131.01927.cbeazley@hgmp.mrc.ac.uk> Is there an agreed consensus on whether or not the optimization flags are set when compiling stuff in EMBOSS? I ask this, because the optimization flags can seriously mess up threaded applications and so knowing whether or not libs have been compiled with optimizations end up being important. claude From gwilliam at hgmp.mrc.ac.uk Fri Nov 15 09:35:35 2002 From: gwilliam at hgmp.mrc.ac.uk (Gary Williams, Tel 01223 494522) Date: Fri, 15 Nov 2002 14:35:35 +0000 Subject: Fw: Other: EMBOSS versus GCG? References: <007e01c28bf3$96beddc0$6886d182@vir.gla.ac.uk> <3DD4F48D.2020606@uk.lionbioscience.com> Message-ID: <3DD50636.8EF0EF45@hgmp.mrc.ac.uk> I started to produce some code for a feature plot, but got bogged down in graphics and have been waiting for the new graphics libraries to appear before continuing. If anyone wants to look at what I have already, I can mail it to them. I would like it to output a graphics image and a file that could be used as an HTML client-side or server-side image active area map. Gary Peter Rice wrote: > > Rory Bowden wrote: > > This on the 'evoldir' list, which is the main international mailing list in > > evolutionary biology. Would anyone like to make any comments for me to pass > > on? > >> It is my impression > >>that GCG is in a different league. For example, compare the GCG program > >>"Window" with its proposed EMBOSS alternative "Freak": > >> > >>TASK: Determination of the number of occurences of a motif in a sequence > >>window. > >> > >>GGC program WINDOW > >> > >>1. Allows up to 6 motifs at a time > >>2. Outputs absolute values and has a variety of other output options. > >>3. Extensive input menu > >> > >>EMBOSS program FREAK > >> > >>1. Allows only 1 motif at a time. > >>2. Outputs a calculated fraction. > >>3. Very limited input menu. > > Window : produces scores over a 'window' (a base range). > StatPlot : Plots Window results > > EMBOSS : reports have scores over a base range as a general output format. > Freak: frequency of matches > FuzzNuc/FuzzPro/FuzzTran: Pattern matches with ambiguity codes > Restrict: Pattern matches with a pattern file > etc... > > This makes it possible to develop some really nice new EMBOSS applications. > > So ... how about a program which reads EMBOSS report files and produces a > summary report (think of window), and another that plots them all (think of > statplot). Scores could be plotted if we have a good way to compare them. > > Yes, I know freak does not produce a report file ... but that is a very > easy change. > > It could also read in EMBL/SwissProt feature tables as annotation. > > So, suggestions please for EMBOSS applications to plot reports/features... > > For example: > > 1. xy plot of scores as points at the centre of a feature, with the > sequence position on the x axis and the score on the y axis. Possibly split > into multiple plots by program/feature-type/named-tag-value (e.g. pattern) > (like statplot only much more versatile). > > 2. xy plot of lines for each feature > > 3. GANTT (bar) chart of features by position, annotated with feature > type/program/score as appropriate > > 4. Combine these - xy plot of features with scores, and other features > reported underneath (think of the -mark option in statplot - but with far > more annotation possible below the x axis) > > Maybe we can make some mock-ups on the EMBOSS pages to show the possibilities? > > regards, > > Peter > > -- > ------------------------------------------------ > Peter Rice, LION Bioscience Ltd, Cambridge, UK > peter.rice at uk.lionbioscience.com +44 1223 224723 -- Gary Williams Tel: +44 1223 494522 Fax: +44 1223 494512 mailto:G.Williams at hgmp.mrc.ac.uk http://www.hgmp.mrc.ac.uk/ Bioinformatics,MRC HGMP Resource Centre,Hinxton,Cambridge, CB10 1SB,UK From root at qnx.com Sat Nov 16 15:10:34 2002 From: root at qnx.com (Jose R. Valverde) Date: Sat, 16 Nov 2002 21:10:34 +0100 (met) Subject: Neural nets Message-ID: The Applications www page states that neural network routines would be a nice add on. So, here's a meagre contribution from my experience with NNs: There are lots of free packages. Rather than coding our own set of routines it may make sense to use one of these. As for NN routines, I'd advise using XNN. It's a GPL environment to build, test and train neural networks. It's got a nice GUI, supports almost all types of networks, and generates the C code routines to be used in the compiled program. Much more effective (IMHO) than coding a library. Other packages exist, but XNN is a nice and complete one. I would like to propose that XNN be used in EMBOSS (say, bundled together) and the NN library be removed from the TO DO list. To get it: ftp://ftp.ii.uib.no/pub/neureka/ For the curious, a starting pointer to many NNs is at http://www.emsl.pnl.gov:2080/proj/neuron/neural/systems/shareware.html from there you may also follow links to freeware for fuzzy systems and genetic algorithms, but I have less experience with these. j From jrvalverde at cnb.uam.es Tue Nov 26 09:43:45 2002 From: jrvalverde at cnb.uam.es (jrvalverde at cnb.uam.es) Date: Tue, 26 Nov 2002 15:43:45 +0100 (MET) Subject: QNX port Message-ID: <200211261443.gAQEhlk2704276@embnet.cnb.uam.es> I've ported EMBOSS to QNX... well, no big deal, it's been rather easy, only a couple of minor changes. Oddly enough the same I had to do for version 2.0x... I'm trying to make the patches to configure, but meanwhile if anybody's interested, just let me know. QNX is a Real Time UNIX operating system used in embedded appliances (e.g. the navigation system in cars and data acquisition systems). The port does not make use of advanced QNX facilities, but works. The binary is available for download as a QNX package to be installed automatically by the QNX packager. The URL is http://www.es.embnet.org/~jr/qnx/ I'm setting up a QNX repository, but it's not totally yet there. j From gwilliam at hgmp.mrc.ac.uk Fri Nov 29 08:58:52 2002 From: gwilliam at hgmp.mrc.ac.uk (gwilliam at hgmp.mrc.ac.uk) Date: Fri, 29 Nov 2002 13:58:52 GMT Subject: GCG Message-ID: <200211291358.gATDwqt25689@californium.hgmp.mrc.ac.uk> Martin Bishop pointed out this GCG site: http://www.hoselton.net/religion/global/split/ :-) Gary From peter.rice at uk.lionbioscience.com Fri Nov 29 09:13:49 2002 From: peter.rice at uk.lionbioscience.com (Peter Rice) Date: Fri, 29 Nov 2002 14:13:49 +0000 Subject: GCG References: <200211291358.gATDwqt25689@californium.hgmp.mrc.ac.uk> Message-ID: <3DE7761D.1080105@uk.lionbioscience.com> gwilliam at hgmp.mrc.ac.uk wrote: > Martin Bishop pointed out this GCG site: > > http://www.hoselton.net/religion/global/split/ From their beliefs - not on that page (as www.gcg.org is no longer theirs :-) but from the UK branch at http://www.globalcog.fsnet.co.uk/ in the "Beliefs" link: The seventh-day weekly Sabbath from Friday sunset to Saturday sunset. The day symbolizes God's resting on the seventh day, after having recreated the surface of the earth in six days, and the forthcoming Millennium which is also referred to as a Sabbath lasting for 1,000 years (Leviticus 23:3; Exodus 20:8-11). So that's why they've not developed anything since January 2000 :-) Peter -- ------------------------------------------------ Peter Rice, LION Bioscience Ltd, Cambridge, UK peter.rice at uk.lionbioscience.com +44 1223 224723 From cbonnard at isrec-sg1.unil.ch Fri Nov 1 09:21:20 2002 From: cbonnard at isrec-sg1.unil.ch (Claude Bonnard) Date: Fri, 1 Nov 2002 10:21:20 +0100 Subject: Remote getz from emboss Message-ID: <10211011021.ZM29739@isrec-sg1> Dear, Does someone use a script (shell or perl) to access a srs server on an other machine with rsh? If yes, is it possible to have this script? the idea is to replace the "local" getz by a rgetz. I am aware of the gbwget development, but I am sure that the equivalent for getz has been done somewhere. -Many thanks in advance, regards, Claude -- Claude Bonnard Ph.D. ISREC (Swiss Institute for Experimental Cancer Research) Bioinformatics Group Ch des Boveresses 155 CH-1066 Epalinges Switzerland phone: [41-21]-692-5891 fax: [41-21]-652-6933 From rls at ebi.ac.uk Fri Nov 1 09:47:17 2002 From: rls at ebi.ac.uk (Rodrigo Lopez) Date: Fri, 1 Nov 2002 09:47:17 -0000 Subject: Remote getz from emboss In-Reply-To: <10211011021.ZM29739@isrec-sg1> Message-ID: Hi, (not directly related to rsh...) :-) The EBI's External Services group is working on a web service that uses SOAP and Axis to address this. We have a prototype running in-house that permits us to use the service within EMBOSS as an external method. Sharmila Pilai, who has been working on this can provide more details. From the user point of view things look like this: % seqret wsembl:hscfos\* -auto -out myfile.tfa The above will use the web service to retrieve data from a source such as SRS, server w/ cdrom index files (another emboss installation), etc. The user needs to install a perl or java client on his/her machine an have some bits and bobs to run the SOAP/Axis client part of the service. Another thing we are working on is being able to use EMBOSS remotely using the same technology, for example from a Windows2000 machine using the command prompt. All that would be required is again a perl/java client and general usage looks as follows: C:\My Documents\wsemboss {appname} {params} Obviously, all can be hidden behind batch scripts to make it even simple to use and in effect, design a simple workflow such as: C:\My Documents\wseqret wsembl:hscfos\* -auto -stdout | wnewcpgreport -filter -auto Anyway, I think Sharmila will have some more comments to make about this and tell you more about her progress. Sharmila, I think you should join this list (if you are not already there?). R:) > -----Original Message----- > From: owner-emboss-dev at hgmp.mrc.ac.uk > [mailto:owner-emboss-dev at hgmp.mrc.ac.uk]On Behalf Of Claude Bonnard > Sent: 01 November 2002 09:21 > To: EMBOSS Dev > Subject: Remote getz from emboss > > > Dear, > > Does someone use a script (shell or perl) to access a srs server > on an other > machine with rsh? If yes, is it possible to have this script? > > the idea is to replace the "local" getz by a rgetz. > I am aware of the gbwget development, but I am sure that the > equivalent for > getz has been done somewhere. > -Many thanks in advance, > > regards, > > Claude > > > -- > Claude Bonnard Ph.D. > ISREC (Swiss Institute for Experimental Cancer Research) > Bioinformatics Group > Ch des Boveresses 155 > CH-1066 Epalinges > Switzerland > phone: [41-21]-692-5891 > fax: [41-21]-652-6933 > From cbeazley at hgmp.mrc.ac.uk Tue Nov 12 11:31:01 2002 From: cbeazley at hgmp.mrc.ac.uk (Claude Beazley) Date: Tue, 12 Nov 2002 11:31:01 +0000 Subject: gcc Optimization of code Message-ID: <200211121131.01927.cbeazley@hgmp.mrc.ac.uk> Is there an agreed consensus on whether or not the optimization flags are set when compiling stuff in EMBOSS? I ask this, because the optimization flags can seriously mess up threaded applications and so knowing whether or not libs have been compiled with optimizations end up being important. claude From gwilliam at hgmp.mrc.ac.uk Fri Nov 15 14:35:35 2002 From: gwilliam at hgmp.mrc.ac.uk (Gary Williams, Tel 01223 494522) Date: Fri, 15 Nov 2002 14:35:35 +0000 Subject: Fw: Other: EMBOSS versus GCG? References: <007e01c28bf3$96beddc0$6886d182@vir.gla.ac.uk> <3DD4F48D.2020606@uk.lionbioscience.com> Message-ID: <3DD50636.8EF0EF45@hgmp.mrc.ac.uk> I started to produce some code for a feature plot, but got bogged down in graphics and have been waiting for the new graphics libraries to appear before continuing. If anyone wants to look at what I have already, I can mail it to them. I would like it to output a graphics image and a file that could be used as an HTML client-side or server-side image active area map. Gary Peter Rice wrote: > > Rory Bowden wrote: > > This on the 'evoldir' list, which is the main international mailing list in > > evolutionary biology. Would anyone like to make any comments for me to pass > > on? > >> It is my impression > >>that GCG is in a different league. For example, compare the GCG program > >>"Window" with its proposed EMBOSS alternative "Freak": > >> > >>TASK: Determination of the number of occurences of a motif in a sequence > >>window. > >> > >>GGC program WINDOW > >> > >>1. Allows up to 6 motifs at a time > >>2. Outputs absolute values and has a variety of other output options. > >>3. Extensive input menu > >> > >>EMBOSS program FREAK > >> > >>1. Allows only 1 motif at a time. > >>2. Outputs a calculated fraction. > >>3. Very limited input menu. > > Window : produces scores over a 'window' (a base range). > StatPlot : Plots Window results > > EMBOSS : reports have scores over a base range as a general output format. > Freak: frequency of matches > FuzzNuc/FuzzPro/FuzzTran: Pattern matches with ambiguity codes > Restrict: Pattern matches with a pattern file > etc... > > This makes it possible to develop some really nice new EMBOSS applications. > > So ... how about a program which reads EMBOSS report files and produces a > summary report (think of window), and another that plots them all (think of > statplot). Scores could be plotted if we have a good way to compare them. > > Yes, I know freak does not produce a report file ... but that is a very > easy change. > > It could also read in EMBL/SwissProt feature tables as annotation. > > So, suggestions please for EMBOSS applications to plot reports/features... > > For example: > > 1. xy plot of scores as points at the centre of a feature, with the > sequence position on the x axis and the score on the y axis. Possibly split > into multiple plots by program/feature-type/named-tag-value (e.g. pattern) > (like statplot only much more versatile). > > 2. xy plot of lines for each feature > > 3. GANTT (bar) chart of features by position, annotated with feature > type/program/score as appropriate > > 4. Combine these - xy plot of features with scores, and other features > reported underneath (think of the -mark option in statplot - but with far > more annotation possible below the x axis) > > Maybe we can make some mock-ups on the EMBOSS pages to show the possibilities? > > regards, > > Peter > > -- > ------------------------------------------------ > Peter Rice, LION Bioscience Ltd, Cambridge, UK > peter.rice at uk.lionbioscience.com +44 1223 224723 -- Gary Williams Tel: +44 1223 494522 Fax: +44 1223 494512 mailto:G.Williams at hgmp.mrc.ac.uk http://www.hgmp.mrc.ac.uk/ Bioinformatics,MRC HGMP Resource Centre,Hinxton,Cambridge, CB10 1SB,UK From root at qnx.com Sat Nov 16 20:10:34 2002 From: root at qnx.com (Jose R. Valverde) Date: Sat, 16 Nov 2002 21:10:34 +0100 (met) Subject: Neural nets Message-ID: The Applications www page states that neural network routines would be a nice add on. So, here's a meagre contribution from my experience with NNs: There are lots of free packages. Rather than coding our own set of routines it may make sense to use one of these. As for NN routines, I'd advise using XNN. It's a GPL environment to build, test and train neural networks. It's got a nice GUI, supports almost all types of networks, and generates the C code routines to be used in the compiled program. Much more effective (IMHO) than coding a library. Other packages exist, but XNN is a nice and complete one. I would like to propose that XNN be used in EMBOSS (say, bundled together) and the NN library be removed from the TO DO list. To get it: ftp://ftp.ii.uib.no/pub/neureka/ For the curious, a starting pointer to many NNs is at http://www.emsl.pnl.gov:2080/proj/neuron/neural/systems/shareware.html from there you may also follow links to freeware for fuzzy systems and genetic algorithms, but I have less experience with these. j From jrvalverde at cnb.uam.es Tue Nov 26 14:43:45 2002 From: jrvalverde at cnb.uam.es (jrvalverde at cnb.uam.es) Date: Tue, 26 Nov 2002 15:43:45 +0100 (MET) Subject: QNX port Message-ID: <200211261443.gAQEhlk2704276@embnet.cnb.uam.es> I've ported EMBOSS to QNX... well, no big deal, it's been rather easy, only a couple of minor changes. Oddly enough the same I had to do for version 2.0x... I'm trying to make the patches to configure, but meanwhile if anybody's interested, just let me know. QNX is a Real Time UNIX operating system used in embedded appliances (e.g. the navigation system in cars and data acquisition systems). The port does not make use of advanced QNX facilities, but works. The binary is available for download as a QNX package to be installed automatically by the QNX packager. The URL is http://www.es.embnet.org/~jr/qnx/ I'm setting up a QNX repository, but it's not totally yet there. j From gwilliam at hgmp.mrc.ac.uk Fri Nov 29 13:58:52 2002 From: gwilliam at hgmp.mrc.ac.uk (gwilliam at hgmp.mrc.ac.uk) Date: Fri, 29 Nov 2002 13:58:52 GMT Subject: GCG Message-ID: <200211291358.gATDwqt25689@californium.hgmp.mrc.ac.uk> Martin Bishop pointed out this GCG site: http://www.hoselton.net/religion/global/split/ :-) Gary From peter.rice at uk.lionbioscience.com Fri Nov 29 14:13:49 2002 From: peter.rice at uk.lionbioscience.com (Peter Rice) Date: Fri, 29 Nov 2002 14:13:49 +0000 Subject: GCG References: <200211291358.gATDwqt25689@californium.hgmp.mrc.ac.uk> Message-ID: <3DE7761D.1080105@uk.lionbioscience.com> gwilliam at hgmp.mrc.ac.uk wrote: > Martin Bishop pointed out this GCG site: > > http://www.hoselton.net/religion/global/split/ From their beliefs - not on that page (as www.gcg.org is no longer theirs :-) but from the UK branch at http://www.globalcog.fsnet.co.uk/ in the "Beliefs" link: The seventh-day weekly Sabbath from Friday sunset to Saturday sunset. The day symbolizes God's resting on the seventh day, after having recreated the surface of the earth in six days, and the forthcoming Millennium which is also referred to as a Sabbath lasting for 1,000 years (Leviticus 23:3; Exodus 20:8-11). So that's why they've not developed anything since January 2000 :-) Peter -- ------------------------------------------------ Peter Rice, LION Bioscience Ltd, Cambridge, UK peter.rice at uk.lionbioscience.com +44 1223 224723