From golharam at umdnj.edu Wed Jan 4 12:05:04 2006 From: golharam at umdnj.edu (Ryan Golhar) Date: Wed, 04 Jan 2006 12:05:04 -0500 Subject: [EMBOSS] Assymetrical Scoring Matrix Message-ID: <018f01c61151$028fcec0$2f01a8c0@GOLHARMOBILE1> I have an assymetrical scoring matrix ie the substitution scores from organism A to organism B is not the same as from organism B to organism A. How can I be sure that when I'm comparing sequence A to sequence B, that stretcher is using the correct side of the diagonal? Ryan From pmr at ebi.ac.uk Thu Jan 5 04:40:30 2006 From: pmr at ebi.ac.uk (pmr at ebi.ac.uk) Date: Thu, 5 Jan 2006 09:40:30 -0000 (GMT) Subject: [EMBOSS] Assymetrical Scoring Matrix In-Reply-To: <018f01c61151$028fcec0$2f01a8c0@GOLHARMOBILE1> References: <018f01c61151$028fcec0$2f01a8c0@GOLHARMOBILE1> Message-ID: <4819.86.137.134.57.1136454030.squirrel@webmail.ebi.ac.uk> Hi Ryan, > I have an assymetrical scoring matrix ie the substitution scores from > organism A to organism B is not the same as from organism B to organism > A. > > How can I be sure that when I'm comparing sequence A to sequence B, that > stretcher is using the correct side of the diagonal? Use two test sequences that have one residue repeated for sequence A and another repeated for sequence B (with a positive score in the matrix) and check the alignment score to see which was used. Then swap the sequences and try again. For example: stretcher asis::qqqqqqqqqq asis:nnnnnnnnnn The alignment score will be 10 times the score that was used (as there are 10 residues in each sequence) Hope that helps ... and this is something we should add to our QA testing. Peter From dag at sonsorol.org Wed Jan 18 17:22:30 2006 From: dag at sonsorol.org (Chris Dagdigian) Date: Wed, 18 Jan 2006 17:22:30 -0500 Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 Message-ID: Hello, Was wondering if anyone else has recently had trouble with 'marscan' in Emboss 2.10 on Apple systems. I'm trying to track down the root cause for someone who last successfully ran marscan on an Apple OS X based system in October 2005. The October run ran just fine. All runs done this week fail silently with zero size output files created and no error messages whatsoever. The only difference is that Apple software updates have been applied to the system. Today, I built 2.10 from source on OS X 10.4 Tiger Server and was able to recreate the silent failure. I then rebuilt 2.10 on my laptop (OS X Tiger 10.4.4) and also have the same silent failure behavior. At this point I'm wondering if something in one of the recent Apple software updates may be the root cause, it really is the only core thing that changed since October. Any hints/suggestions? Can others reproduce? Regards, Chris Dagdigian From kvddrift at earthlink.net Wed Jan 18 18:04:53 2006 From: kvddrift at earthlink.net (Koen van der Drift) Date: Wed, 18 Jan 2006 18:04:53 -0500 Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 In-Reply-To: References: Message-ID: On Jan 18, 2006, at 5:22 PM, Chris Dagdigian wrote: > > Hello, > > Was wondering if anyone else has recently had trouble with 'marscan' > in Emboss 2.10 on Apple systems. > > I'm trying to track down the root cause for someone who last > successfully ran marscan on an Apple OS X based system in October > 2005. > > The October run ran just fine. All runs done this week fail silently > with zero size output files created and no error messages whatsoever. > > The only difference is that Apple software updates have been applied > to the system. > > Today, I built 2.10 from source on OS X 10.4 Tiger Server and was > able to recreate the silent failure. > > I then rebuilt 2.10 on my laptop (OS X Tiger 10.4.4) and also have > the same silent failure behavior. > > At this point I'm wondering if something in one of the recent Apple > software updates may be the root cause, it really is the only core > thing that changed since October. > > Any hints/suggestions? Can others reproduce? > > Regards, > Chris Dagdigian I'm getting a similar error on 10.4.4, using EMBOSS 3.0.0, installed using fink. I've never used marscan before, but fed it a fasta file to test, and the output indeed is empty. - Koen. From dag at sonsorol.org Fri Jan 20 15:33:14 2006 From: dag at sonsorol.org (Chris Dagdigian) Date: Fri, 20 Jan 2006 15:33:14 -0500 Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 In-Reply-To: <3920.86.134.65.23.1137672316.squirrel@webmail.ebi.ac.uk> References: <53936.81.96.70.96.1137631247.squirrel@webmail.ebi.ac.uk> <36578.81.96.70.96.1137671293.squirrel@webmail.ebi.ac.uk> <3920.86.134.65.23.1137672316.squirrel@webmail.ebi.ac.uk> Message-ID: <4BFED776-CBCD-4946-B0C9-0F8FEB7A91B9@sonsorol.org> Thanks for all the advice but I've still got marscan failing silently on me with Emboss 2.10 on OS X. The given EMBOSS test sequence hsbhh.fasta works perfectly fine. My other test case containing sequence given to me still generates zero size output files with no content. The 'failing' input is a normal fasta-formatted DNA sequence roughly 2000 bp with no special or odd characters or alphabet members. So far all file twiddling efforts via BBedit and emacs have had no effect: - changing from upper to lower case - adding/removing the fasta naming line - adding/removing line breaks - changing line wrapping from 0 to 72 and 50 chars - DOS linebreaks - Unix linebreaks - Mac linebreaks - Windows Latin 1 encoding - Mac Roman encoding I'm stumped. Either something really simple is wrong with my query or perhaps marscan is just not writing any output if it does not find anything? I'm going to see if its OK to post the query sequence in this forum. Regards, Chris On Jan 19, 2006, at 7:05 AM, pmr at ebi.ac.uk wrote: > Alan writes, > >> There used to be a problem with the old Mac end-of-line characters >> but >> I believe that was solved before 2.10.0. It may be worth checking >> precisely what format your fasta file is in though: EMBOSS has always >> been OK with just standard UNIX terminators. > > Yup, that would be something that could cause an empty sequence - if > EMBOSS fails to find some kind of end-of-line character then the > seqeunce > becomes just part of the description in FASTA format. > > Does editing the input file make a difference? > > Does seqret like the same input sequence? > > Hope that helps, > > Peter From pmr at ebi.ac.uk Sat Jan 21 07:58:50 2006 From: pmr at ebi.ac.uk (pmr at ebi.ac.uk) Date: Sat, 21 Jan 2006 12:58:50 -0000 (GMT) Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 In-Reply-To: <4BFED776-CBCD-4946-B0C9-0F8FEB7A91B9@sonsorol.org> References: <53936.81.96.70.96.1137631247.squirrel@webmail.ebi.ac.uk> <36578.81.96.70.96.1137671293.squirrel@webmail.ebi.ac.uk> <3920.86.134.65.23.1137672316.squirrel@webmail.ebi.ac.uk> <4BFED776-CBCD-4946-B0C9-0F8FEB7A91B9@sonsorol.org> Message-ID: <4901.86.133.36.229.1137848330.squirrel@webmail.ebi.ac.uk> Hi Chris, > I'm going to see if its OK to post the query sequence in this forum. Or you could run with -debug on the command line and send the marscan.dbg file to emboss-bug at emboss.open-bio.org regards, Peter Rice From Marc.Logghe at DEVGEN.com Mon Jan 23 04:52:14 2006 From: Marc.Logghe at DEVGEN.com (Marc Logghe) Date: Mon, 23 Jan 2006 10:52:14 +0100 Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 Message-ID: <0C528E3670D8CE4B8E013F6749231AA6746A70@ANTARESIA.be.devgen.com> Hi Chris, Do you see any particularities when you compare the working sequence with the non working ones in debug mode ? Cheers, Marc > Thanks for all the advice but I've still got marscan failing > silently on me with Emboss 2.10 on OS X. > > The given EMBOSS test sequence hsbhh.fasta works perfectly fine. > > My other test case containing sequence given to me still > generates zero size output files with no content. > > The 'failing' input is a normal fasta-formatted DNA sequence > roughly 2000 bp with no special or odd characters or alphabet members. > > So far all file twiddling efforts via BBedit and emacs have had no > effect: > > - changing from upper to lower case > - adding/removing the fasta naming line > - adding/removing line breaks > - changing line wrapping from 0 to 72 and 50 chars > - DOS linebreaks > - Unix linebreaks > - Mac linebreaks > - Windows Latin 1 encoding > - Mac Roman encoding > > I'm stumped. Either something really simple is wrong with my > query or perhaps marscan is just not writing any output if it > does not find anything? > > I'm going to see if its OK to post the query sequence in this forum. > > Regards, > Chris > > > > > On Jan 19, 2006, at 7:05 AM, pmr at ebi.ac.uk wrote: > > > Alan writes, > > > >> There used to be a problem with the old Mac end-of-line characters > >> but I believe that was solved before 2.10.0. It may be > worth checking > >> precisely what format your fasta file is in though: EMBOSS > has always > >> been OK with just standard UNIX terminators. > > > > Yup, that would be something that could cause an empty > sequence - if > > EMBOSS fails to find some kind of end-of-line character then the > > seqeunce becomes just part of the description in FASTA format. > > > > Does editing the input file make a difference? > > > > Does seqret like the same input sequence? > > > > Hope that helps, > > > > Peter > > _______________________________________________ > EMBOSS mailing list > EMBOSS at emboss.open-bio.org > http://newportal.open-bio.org/mailman/listinfo/emboss > From ajb at ebi.ac.uk Mon Jan 23 07:12:21 2006 From: ajb at ebi.ac.uk (ajb at ebi.ac.uk) Date: Mon, 23 Jan 2006 12:12:21 -0000 (GMT) Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 In-Reply-To: <0C528E3670D8CE4B8E013F6749231AA6746A70@ANTARESIA.be.devgen.com> References: <0C528E3670D8CE4B8E013F6749231AA6746A70@ANTARESIA.be.devgen.com> Message-ID: <35706.172.31.100.157.1138018341.squirrel@webmail.ebi.ac.uk> Marscan was, confusingly, not sending suitable negative results to an output file for some test sequences with no matching sites (and just producing an empty output file). This was happening on all OSs. A fix has been put into: ftp://emboss.open-bio.org/pub/EMBOSS/fixes/marscan.c This will work on versions 3 & 2.10 Alan From golharam at umdnj.edu Wed Jan 4 17:05:04 2006 From: golharam at umdnj.edu (Ryan Golhar) Date: Wed, 04 Jan 2006 12:05:04 -0500 Subject: [EMBOSS] Assymetrical Scoring Matrix Message-ID: <018f01c61151$028fcec0$2f01a8c0@GOLHARMOBILE1> I have an assymetrical scoring matrix ie the substitution scores from organism A to organism B is not the same as from organism B to organism A. How can I be sure that when I'm comparing sequence A to sequence B, that stretcher is using the correct side of the diagonal? Ryan From pmr at ebi.ac.uk Thu Jan 5 09:40:30 2006 From: pmr at ebi.ac.uk (pmr at ebi.ac.uk) Date: Thu, 5 Jan 2006 09:40:30 -0000 (GMT) Subject: [EMBOSS] Assymetrical Scoring Matrix In-Reply-To: <018f01c61151$028fcec0$2f01a8c0@GOLHARMOBILE1> References: <018f01c61151$028fcec0$2f01a8c0@GOLHARMOBILE1> Message-ID: <4819.86.137.134.57.1136454030.squirrel@webmail.ebi.ac.uk> Hi Ryan, > I have an assymetrical scoring matrix ie the substitution scores from > organism A to organism B is not the same as from organism B to organism > A. > > How can I be sure that when I'm comparing sequence A to sequence B, that > stretcher is using the correct side of the diagonal? Use two test sequences that have one residue repeated for sequence A and another repeated for sequence B (with a positive score in the matrix) and check the alignment score to see which was used. Then swap the sequences and try again. For example: stretcher asis::qqqqqqqqqq asis:nnnnnnnnnn The alignment score will be 10 times the score that was used (as there are 10 residues in each sequence) Hope that helps ... and this is something we should add to our QA testing. Peter From dag at sonsorol.org Wed Jan 18 22:22:30 2006 From: dag at sonsorol.org (Chris Dagdigian) Date: Wed, 18 Jan 2006 17:22:30 -0500 Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 Message-ID: Hello, Was wondering if anyone else has recently had trouble with 'marscan' in Emboss 2.10 on Apple systems. I'm trying to track down the root cause for someone who last successfully ran marscan on an Apple OS X based system in October 2005. The October run ran just fine. All runs done this week fail silently with zero size output files created and no error messages whatsoever. The only difference is that Apple software updates have been applied to the system. Today, I built 2.10 from source on OS X 10.4 Tiger Server and was able to recreate the silent failure. I then rebuilt 2.10 on my laptop (OS X Tiger 10.4.4) and also have the same silent failure behavior. At this point I'm wondering if something in one of the recent Apple software updates may be the root cause, it really is the only core thing that changed since October. Any hints/suggestions? Can others reproduce? Regards, Chris Dagdigian From kvddrift at earthlink.net Wed Jan 18 23:04:53 2006 From: kvddrift at earthlink.net (Koen van der Drift) Date: Wed, 18 Jan 2006 18:04:53 -0500 Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 In-Reply-To: References: Message-ID: On Jan 18, 2006, at 5:22 PM, Chris Dagdigian wrote: > > Hello, > > Was wondering if anyone else has recently had trouble with 'marscan' > in Emboss 2.10 on Apple systems. > > I'm trying to track down the root cause for someone who last > successfully ran marscan on an Apple OS X based system in October > 2005. > > The October run ran just fine. All runs done this week fail silently > with zero size output files created and no error messages whatsoever. > > The only difference is that Apple software updates have been applied > to the system. > > Today, I built 2.10 from source on OS X 10.4 Tiger Server and was > able to recreate the silent failure. > > I then rebuilt 2.10 on my laptop (OS X Tiger 10.4.4) and also have > the same silent failure behavior. > > At this point I'm wondering if something in one of the recent Apple > software updates may be the root cause, it really is the only core > thing that changed since October. > > Any hints/suggestions? Can others reproduce? > > Regards, > Chris Dagdigian I'm getting a similar error on 10.4.4, using EMBOSS 3.0.0, installed using fink. I've never used marscan before, but fed it a fasta file to test, and the output indeed is empty. - Koen. From dag at sonsorol.org Fri Jan 20 20:33:14 2006 From: dag at sonsorol.org (Chris Dagdigian) Date: Fri, 20 Jan 2006 15:33:14 -0500 Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 In-Reply-To: <3920.86.134.65.23.1137672316.squirrel@webmail.ebi.ac.uk> References: <53936.81.96.70.96.1137631247.squirrel@webmail.ebi.ac.uk> <36578.81.96.70.96.1137671293.squirrel@webmail.ebi.ac.uk> <3920.86.134.65.23.1137672316.squirrel@webmail.ebi.ac.uk> Message-ID: <4BFED776-CBCD-4946-B0C9-0F8FEB7A91B9@sonsorol.org> Thanks for all the advice but I've still got marscan failing silently on me with Emboss 2.10 on OS X. The given EMBOSS test sequence hsbhh.fasta works perfectly fine. My other test case containing sequence given to me still generates zero size output files with no content. The 'failing' input is a normal fasta-formatted DNA sequence roughly 2000 bp with no special or odd characters or alphabet members. So far all file twiddling efforts via BBedit and emacs have had no effect: - changing from upper to lower case - adding/removing the fasta naming line - adding/removing line breaks - changing line wrapping from 0 to 72 and 50 chars - DOS linebreaks - Unix linebreaks - Mac linebreaks - Windows Latin 1 encoding - Mac Roman encoding I'm stumped. Either something really simple is wrong with my query or perhaps marscan is just not writing any output if it does not find anything? I'm going to see if its OK to post the query sequence in this forum. Regards, Chris On Jan 19, 2006, at 7:05 AM, pmr at ebi.ac.uk wrote: > Alan writes, > >> There used to be a problem with the old Mac end-of-line characters >> but >> I believe that was solved before 2.10.0. It may be worth checking >> precisely what format your fasta file is in though: EMBOSS has always >> been OK with just standard UNIX terminators. > > Yup, that would be something that could cause an empty sequence - if > EMBOSS fails to find some kind of end-of-line character then the > seqeunce > becomes just part of the description in FASTA format. > > Does editing the input file make a difference? > > Does seqret like the same input sequence? > > Hope that helps, > > Peter From pmr at ebi.ac.uk Sat Jan 21 12:58:50 2006 From: pmr at ebi.ac.uk (pmr at ebi.ac.uk) Date: Sat, 21 Jan 2006 12:58:50 -0000 (GMT) Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 In-Reply-To: <4BFED776-CBCD-4946-B0C9-0F8FEB7A91B9@sonsorol.org> References: <53936.81.96.70.96.1137631247.squirrel@webmail.ebi.ac.uk> <36578.81.96.70.96.1137671293.squirrel@webmail.ebi.ac.uk> <3920.86.134.65.23.1137672316.squirrel@webmail.ebi.ac.uk> <4BFED776-CBCD-4946-B0C9-0F8FEB7A91B9@sonsorol.org> Message-ID: <4901.86.133.36.229.1137848330.squirrel@webmail.ebi.ac.uk> Hi Chris, > I'm going to see if its OK to post the query sequence in this forum. Or you could run with -debug on the command line and send the marscan.dbg file to emboss-bug at emboss.open-bio.org regards, Peter Rice From Marc.Logghe at DEVGEN.com Mon Jan 23 09:52:14 2006 From: Marc.Logghe at DEVGEN.com (Marc Logghe) Date: Mon, 23 Jan 2006 10:52:14 +0100 Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 Message-ID: <0C528E3670D8CE4B8E013F6749231AA6746A70@ANTARESIA.be.devgen.com> Hi Chris, Do you see any particularities when you compare the working sequence with the non working ones in debug mode ? Cheers, Marc > Thanks for all the advice but I've still got marscan failing > silently on me with Emboss 2.10 on OS X. > > The given EMBOSS test sequence hsbhh.fasta works perfectly fine. > > My other test case containing sequence given to me still > generates zero size output files with no content. > > The 'failing' input is a normal fasta-formatted DNA sequence > roughly 2000 bp with no special or odd characters or alphabet members. > > So far all file twiddling efforts via BBedit and emacs have had no > effect: > > - changing from upper to lower case > - adding/removing the fasta naming line > - adding/removing line breaks > - changing line wrapping from 0 to 72 and 50 chars > - DOS linebreaks > - Unix linebreaks > - Mac linebreaks > - Windows Latin 1 encoding > - Mac Roman encoding > > I'm stumped. Either something really simple is wrong with my > query or perhaps marscan is just not writing any output if it > does not find anything? > > I'm going to see if its OK to post the query sequence in this forum. > > Regards, > Chris > > > > > On Jan 19, 2006, at 7:05 AM, pmr at ebi.ac.uk wrote: > > > Alan writes, > > > >> There used to be a problem with the old Mac end-of-line characters > >> but I believe that was solved before 2.10.0. It may be > worth checking > >> precisely what format your fasta file is in though: EMBOSS > has always > >> been OK with just standard UNIX terminators. > > > > Yup, that would be something that could cause an empty > sequence - if > > EMBOSS fails to find some kind of end-of-line character then the > > seqeunce becomes just part of the description in FASTA format. > > > > Does editing the input file make a difference? > > > > Does seqret like the same input sequence? > > > > Hope that helps, > > > > Peter > > _______________________________________________ > EMBOSS mailing list > EMBOSS at emboss.open-bio.org > http://newportal.open-bio.org/mailman/listinfo/emboss > From ajb at ebi.ac.uk Mon Jan 23 12:12:21 2006 From: ajb at ebi.ac.uk (ajb at ebi.ac.uk) Date: Mon, 23 Jan 2006 12:12:21 -0000 (GMT) Subject: [EMBOSS] emboss 2.10 'marscan' now silently fails on Mac OS X 10.4 In-Reply-To: <0C528E3670D8CE4B8E013F6749231AA6746A70@ANTARESIA.be.devgen.com> References: <0C528E3670D8CE4B8E013F6749231AA6746A70@ANTARESIA.be.devgen.com> Message-ID: <35706.172.31.100.157.1138018341.squirrel@webmail.ebi.ac.uk> Marscan was, confusingly, not sending suitable negative results to an output file for some test sequences with no matching sites (and just producing an empty output file). This was happening on all OSs. A fix has been put into: ftp://emboss.open-bio.org/pub/EMBOSS/fixes/marscan.c This will work on versions 3 & 2.10 Alan