[maker-devel] Using maker with precomputed transcript / protein alignments
Carson Holt
carsonhh at gmail.com
Fri May 24 14:06:51 MDT 2013
I'm glad it's working. I think I'll add a check for the '/' characters in
the base name as I think having it be a directory will get me in trouble
somewhere with hidden bugs.
Thanks,
Carson
From: Daniel Standage <daniel.standage at gmail.com>
Date: Friday, 24 May, 2013 4:00 PM
To: Carson Holt <carsonhh at gmail.com>
Subject: Re: [maker-devel] Using maker with precomputed transcript /
protein alignments
Oh wow, you are going to LOVE this.
I kept on messing around with things to see if I could tease out any
patterns, and eventually it hit me. In my working directory, I have an
outputs directory, which is intended to contain output directories from
various different maker runs. However, since my submission scripts launch
Maker from the working directory, I use -base outputs/blahblahblah as the
base parameter. So when it tries to create output files using the base name
(the SQLite3 db just happens to be the first), it tries to create
outputs/blahblahblah/outputs/blahblahblah.db, and of course that internal
outputs directory doesn't exist. Every time I've had problems, I've been
using a basename with a / character (relative directory path). Every time I
haven't had problems, it was because the / wasn't there.
Since the base parameter determines the name of the output directory, I
assumed I could also use it specify a nested output directory. So it looks
like I just need to be more careful that the basenames I use don't contain /
characters or any other special UNIX characters. Of course, this could be
made explicit in the usage statement, or you could add something like this
right after parsing the command line arguments.
if($OPT{"out_name"} =~ m/\//)
{
printf(STDERR "base '%s' invalid: basenames containing relative directory
paths cause errors; please provide a simple string instead",
$OPT{"out_name"});
exit_maker(0);
}
Alternatively, you could handle things like I had originally expected: if I
provide path/to/mybase as my base parameter, maker would create the
path/to/mybase directory initially, but then in the creation of subsequent
files it would simply use mybase. I don't imagine this would be too
extensive of a change, but I understand Maker has a huge codebase. Anyway,
just some suggestions, take them for what they're worth.
Thanks for your help!
--
Daniel S. Standage
Ph.D. Candidate
Bioinformatics and Computational Biology Program
Department of Genetics, Development, and Cell Biology
Iowa State University
On Fri, May 24, 2013 at 3:29 PM, Carson Holt <carsonhh at gmail.com> wrote:
> NFS is weird. It's hard to say why it was freezing the first times, and did
> not appear to freeze on your very last try. I definitely want to know if it
> starts to freeze again, or if stack traces show a consistent point where it
> freezes. If it keeps happening, I can try making the database in the local
> /tmp and then just copying it to the current working directory once it's
> populated to get around any weird NFS issues. But before going through all
> the effort to do that, I'd like to know that it's not some other weird bug
> related to the perl your using or other modules that are installed. Top
> candidates on the list would be modules such as forks, forks::shared, DBI, or
> DBD::SQLite. Try reinstalling those
>
> Thanks,
> Carson
>
>
> From: Daniel Standage <daniel.standage at gmail.com>
> Date: Friday, 24 May, 2013 3:19 PM
>
> To: Carson Holt <carsonhh at gmail.com>
> Subject: Re: [maker-devel] Using maker with precomputed transcript / protein
> alignments
>
> I admit I killed these last few runs too quickly, I guess I was getting
> impatient, especially since waiting hours or days hasn't made a difference
> before. Either way, that was sloppy on my part.
>
> However, I always specify the base parameter, whether or not I'm running
> mulitple maker jobs from the same directory. And if I ever restarted a job, I
> have always removed the original output directory entirely before
> relaunching--precisely to avoid the types of mistakes you mention arising from
> residual files.
>
> --
> Daniel S. Standage
> Ph.D. Candidate
> Bioinformatics and Computational Biology Program
> Department of Genetics, Development, and Cell Biology
> Iowa State University
>
>
> On Fri, May 24, 2013 at 3:10 PM, Carson Holt <carsonhh at gmail.com> wrote:
>> Correct if you use the -base parameter you should get a different output
>> directory. And if you have never used that base before, and it still
>> freezes, then there is a problem. You do need to give it a little more time
>> until killing it, as the stack trace in both cases showed that it was less
>> than 25% finished reading the input GFF3 files and even less than that in the
>> first case (so give it about 5x as long before giving up).
>>
>> It might just be that the NFS mount is slow. Or because of how weird the
>> error is, other options include reinstalling perl and all modules. The
>> weirdest bugs are often broken perl or inadvertently using modules from
>> different perl versions via the PERL5LIB environmental variable (this is very
>> common and can cause very wacky behavior). Another option is verifying all
>> software for the lustre NFS mount is up to date. Lastly there was an odd NFS
>> bug that came up on the e-mail list last week that was fixed by a kernel
>> upgrade.
>>
>> --Carson
>>
>>
>>
>> From: Daniel Standage <daniel.standage at gmail.com>
>> Date: Friday, 24 May, 2013 3:01 PM
>>
>> To: Carson Holt <carsonhh at gmail.com>
>> Subject: Re: [maker-devel] Using maker with precomputed transcript / protein
>> alignments
>>
>> The file locks are created only in the output directory, no? So there is a
>> problem if I have multiple maker runs launched from the same directory, but
>> writing to different output directories (as specified by different base
>> parameters)?
>>
>>
>> --
>> Daniel S. Standage
>> Ph.D. Candidate
>> Bioinformatics and Computational Biology Program
>> Department of Genetics, Development, and Cell Biology
>> Iowa State University
>>
>>
>> On Fri, May 24, 2013 at 2:57 PM, Carson Holt <carsonhh at gmail.com> wrote:
>>> To clarify, that means you need to use a different working directory. Can
>>> be a subdirectory of your original.
>>>
>>> --Carson
>>>
>>>
>>> From: Carson Holt <carsonhh at gmail.com>
>>> Date: Friday, 24 May, 2013 2:56 PM
>>> To: Daniel Standage <daniel.standage at gmail.com>
>>>
>>> Subject: Re: [maker-devel] Using maker with precomputed transcript /
>>> protein alignments
>>>
>>> Both stack traces show different locations in the code and file being read.
>>> So it appears it was not frozen, just interrupted by control-C.
>>>
>>> If you restart make sure you do so in a completely new directory from the
>>> original run. This is because I wonder if there is a failed job that still
>>> has active processes and is holding onto file locks in that directory.
>>>
>>> --Carson
>>>
>>>
>>> From: Daniel Standage <daniel.standage at gmail.com>
>>> Date: Friday, 24 May, 2013 2:50 PM
>>> To: Carson Holt <carsonhh at gmail.com>
>>> Subject: Re: [maker-devel] Using maker with precomputed transcript /
>>> protein alignments
>>>
>>> Deleted output directory and re-ran. Stack trace looks pretty similar.
>>>
>>>
>>> Calling GFFDB::new at /N/u/dstandag/Mason/local/src/maker-dev/bin/maker line
>>> 607.
>>> SIGINT received
>>> at /N/u/dstandag/Mason/local/src/PerlLibs/lib64/perl5/forks/signals.pm
>>> <http://signals.pm> line 97, <$IN
>>>> > line 243676.
>>> forks::signals::__ANON__('INT') called at /usr/lib64/perl5/DBI.pm
>>> line 1590
>>> eval {...} called at /usr/lib64/perl5/DBI.pm line 1590
>>> DBD::_::db::do('DBI::db=HASH(0x4987228)', 'INSERT INTO est_gff
>>> (seqid, source, parent, start, end, line)...') called at
>>> /N/hd01/dstandag/Mason/local/src/maker-dev/bin/../lib/GFFDB.pm line 493
>>> GFFDB::_add_to_db('GFFDB=HASH(0x49727a0)',
>>> 'DBI::db=HASH(0x49871e0)', 'est_gff', 'HASH(0x49877e0)') called at
>>> /N/hd01/dstandag/Mason/local/src/maker-dev/bin/../lib/GFFDB.pm line 432
>>> GFFDB::_add_type('GFFDB=HASH(0x49727a0)',
>>> '/N/dc/scratch/dstandag/PdomGenomic/Annotation/annot-v0.41/inp...',
>>> 'est_gff') called at
>>> /N/hd01/dstandag/Mason/local/src/maker-dev/bin/../lib/GFFDB.pm line 324
>>> GFFDB::add_est('GFFDB=HASH(0x49727a0)',
>>> '/N/dc/scratch/dstandag/PdomGenomic/Annotation/annot-v0.41/inp...') called
>>> at /N/hd01/dstandag/Mason/local/src/maker-dev/bin/../lib/GFFDB.pm line 57
>>> GFFDB::new('GFFDB', 'HASH(0x489c488)') called at
>>> /N/u/dstandag/Mason/local/src/maker-dev/bin/maker line 608
>>>
>>>
>>> --
>>> Daniel S. Standage
>>> Ph.D. Candidate
>>> Bioinformatics and Computational Biology Program
>>> Department of Genetics, Development, and Cell Biology
>>> Iowa State University
>>>
>>>
>>> On Fri, May 24, 2013 at 2:45 PM, Carson Holt <carsonhh at gmail.com> wrote:
>>>> Could you run again, and so I can see if the stack trace is the same each
>>>> time.
>>>>
>>>> --Carson
>>>>
>>>>
>>>> From: Daniel Standage <daniel.standage at gmail.com>
>>>> Date: Friday, 24 May, 2013 2:39 PM
>>>>
>>>> To: Carson Holt <carsonhh at gmail.com>
>>>> Subject: Re: [maker-devel] Using maker with precomputed transcript /
>>>> protein alignments
>>>>
>>>> Restarted in the original NSF-mounted directory, never saw the .db file,
>>>> and got this as the stack trace upon termination.
>>>>
>>>> STATUS: Setting up database for any GFF3 input...
>>>> Calling GFFDB::new at /N/u/dstandag/Mason/local/src/maker-dev/bin/maker
>>>> line 607.
>>>> SIGINT received
>>>> at /N/u/dstandag/Mason/local/src/PerlLibs/lib64/perl5/forks/signals.pm
>>>> <http://signals.pm> line 97, <$IN> line 170294.
>>>> forks::signals::__ANON__('INT') called at
>>>> /N/hd01/dstandag/Mason/local/src/maker-dev/bin/../lib/GFFDB.pm line 475
>>>> eval {...} called at
>>>> /N/hd01/dstandag/Mason/local/src/maker-dev/bin/../lib/GFFDB.pm line 475
>>>> GFFDB::_parse_line('GFFDB=HASH(0x4e5c730)', 'SCALAR(0x4e714b8)',
>>>> 'est_gff') called at
>>>> /N/hd01/dstandag/Mason/local/src/maker-dev/bin/../lib/GFFDB.pm line 431
>>>> GFFDB::_add_type('GFFDB=HASH(0x4e5c730)',
>>>> '/N/dc/scratch/dstandag/PdomGenomic/Annotation/annot-v0.41/inp...',
>>>> 'est_gff') called at
>>>> /N/hd01/dstandag/Mason/local/src/maker-dev/bin/../lib/GFFDB.pm line 324
>>>> GFFDB::add_est('GFFDB=HASH(0x4e5c730)',
>>>> '/N/dc/scratch/dstandag/PdomGenomic/Annotation/annot-v0.41/inp...') called
>>>> at /N/hd01/dstandag/Mason/local/src/maker-dev/bin/../lib/GFFDB.pm line 57
>>>> GFFDB::new('GFFDB', 'HASH(0x4d86488)') called at
>>>> /N/u/dstandag/Mason/local/src/maker-dev/bin/maker line 608
>>>>
>>>>
>>>> --
>>>> Daniel S. Standage
>>>> Ph.D. Candidate
>>>> Bioinformatics and Computational Biology Program
>>>> Department of Genetics, Development, and Cell Biology
>>>> Iowa State University
>>>>
>>>>
>>>> On Fri, May 24, 2013 at 2:25 PM, Carson Holt <carsonhh at gmail.com> wrote:
>>>>> Start a new job in a new directory from the original job (NFS mount). Use
>>>>> the new maker executable I sent. If it still freezes, hit control-C to
>>>>> get a stack trace.
>>>>>
>>>>> --Carson
>>>>>
>>>>>
>>>>> From: Daniel Standage <daniel.standage at gmail.com>
>>>>> Date: Friday, 24 May, 2013 2:21 PM
>>>>>
>>>>> To: Carson Holt <carsonhh at gmail.com>
>>>>> Subject: Re: [maker-devel] Using maker with precomputed transcript /
>>>>> protein alignments
>>>>>
>>>>> The job from several hours ago is still running with no changes.
>>>>>
>>>>> I just relaunched the job with a locally mounted working directory: I
>>>>> could see the .db file almost immediately, and it took less than 5 minutes
>>>>> to successfully build the SQLite3 db and proceed to the next steps of the
>>>>> pipeline. Any ideas?
>>>>>
>>>>> --
>>>>> Daniel S. Standage
>>>>> Ph.D. Candidate
>>>>> Bioinformatics and Computational Biology Program
>>>>> Department of Genetics, Development, and Cell Biology
>>>>> Iowa State University
>>>>>
>>>>>
>>>>> On Fri, May 24, 2013 at 2:01 PM, Carson Holt <carsonhh at gmail.com> wrote:
>>>>>> The NFS mount appears to be configured correctly.
>>>>>>
>>>>>> Here is what the maker.output directory should look like while the
>>>>>> database is being generated.
>>>>>>
>>>>>> drwxr-xr-x 10 cholt staff 340 24 May 13:51 .
>>>>>> drwxr-xr-x 10 cholt staff 340 24 May 13:50 ..
>>>>>> -rw------x 1 cholt staff 85 24 May 13:50
>>>>>> .NFSLock.gi_lock.NFSLock
>>>>>> -rw------- 1 cholt staff 52 24 May 13:50
>>>>>> .NFSLock.pdom-annot-v0.41-1.db.NFSLock
>>>>>> -rw-r--r-- 1 cholt staff 1413 24 May 13:50 maker_bopts.log
>>>>>> -rw-r--r-- 1 cholt staff 1666 24 May 13:50 maker_exe.log
>>>>>> -rw-r--r-- 1 cholt staff 4610 24 May 13:50 maker_opts.log
>>>>>> drwxr-xr-x 4 cholt staff 136 24 May 13:50 mpi_blastdb
>>>>>> -rw-r--r-- 1 cholt staff 29326336 24 May 13:51 pdom-annot-v0.41-1.db
>>>>>> -rw-r--r-- 1 cholt staff 6704 24 May 13:51
>>>>>> pdom-annot-v0.41-1.db-journal
>>>>>>
>>>>>>
>>>>>> Could you watch while maker is running to see if this file is created -->
>>>>>> .NFSLock.pdom-annot-v0.41-1.db.NFSLock
>>>>>> You must use ls with the -a flag to see it or it will be hidden.
>>>>>>
>>>>>> Just keep letting it run until that file shows up. Shortly after it sows
>>>>>> up, this one should appear --> pdom-annot-v0.41-1.db-journal
>>>>>>
>>>>>> Also could you try running MAKER once with the working directory being
>>>>>> locally mounted (/tmp for example).
>>>>>>
>>>>>> --Carson
>>>>>>
>>>>>>
>>>>>>
>>>>>> From: Daniel Standage <daniel.standage at gmail.com>
>>>>>> Date: Friday, 24 May, 2013 1:36 PM
>>>>>>
>>>>>> To: Carson Holt <carsonhh at gmail.com>
>>>>>> Subject: Re: [maker-devel] Using maker with precomputed transcript /
>>>>>> protein alignments
>>>>>>
>>>>>> Here is the output.
>>>>>>
>>>>>> [dstandag at mason annot-v0.41] ls -al
>>>>>> outputs/pdom-annot-v0.41-1.maker.output/
>>>>>> total 32
>>>>>> drwxr-xr-x 3 dstandag biol 4096 May 24 13:34 .
>>>>>> drwxr-xr-x 3 dstandag biol 4096 May 24 12:39 ..
>>>>>> -rw-r--r-- 1 dstandag biol 1413 May 24 12:39 maker_bopts.log
>>>>>> -rw-r--r-- 1 dstandag biol 1355 May 24 12:39 maker_exe.log
>>>>>> -rw-r--r-- 1 dstandag biol 4883 May 24 12:39 maker_opts.log
>>>>>> drwxr-xr-x 3 dstandag biol 4096 May 24 12:39 mpi_blastdb
>>>>>> -rw------x 1 dstandag biol 70 May 24 13:34 .NFSLock.gi_lock.NFSLock
>>>>>> [dstandag at mason annot-v0.41] df outputs/pdom-annot-v0.41-1.maker.output/
>>>>>> Filesystem 1K-blocks Used Available Use% Mounted on
>>>>>> dc-mds01.uits.indiana.edu:/dc
>>>>>> 1144318908992 928977247792 203869022296 83% /N/dc
>>>>>> [dstandag at mason annot-v0.41] mount
>>>>>> login_x86_64 on / type tmpfs (rw)
>>>>>> proc on /proc type proc (rw)
>>>>>> sysfs on /sys type sysfs (rw)
>>>>>> devpts on /dev/pts type devpts (rw,gid=5,mode=620)
>>>>>> tmpfs on /dev/shm type tmpfs (rw)
>>>>>> tmpfs on /var/tmp type tmpfs (rw,size=10m)
>>>>>> /dev/sdb2 on /tmp type ext4 (rw,relatime,barrier=1,data=ordered)
>>>>>> none on /proc/sys/fs/binfmt_misc type binfmt_misc (rw)
>>>>>> sunrpc on /var/lib/nfs/rpc_pipefs type rpc_pipefs (rw)
>>>>>> AFS on /afs type afs (rw)
>>>>>> bl-nas1:/vol/hd00 on /N/hd00 type nfs
>>>>>> (rw,nosuid,tcp,rsize=32768,wsize=32768,timeo=600,retrans=2,intr,addr=149.
>>>>>> 165.226.129)
>>>>>> bl-nas1:/vol/hd01 on /N/hd01 type nfs
>>>>>> (rw,nosuid,tcp,rsize=32768,wsize=32768,timeo=600,retrans=2,intr,addr=149.
>>>>>> 165.226.129)
>>>>>> bl-nas2:/vol/hd02 on /N/hd02 type nfs
>>>>>> (rw,nosuid,tcp,rsize=32768,wsize=32768,timeo=600,retrans=2,intr,addr=149.
>>>>>> 165.226.130)
>>>>>> bl-nas2:/vol/hd03 on /N/hd03 type nfs
>>>>>> (rw,nosuid,tcp,rsize=32768,wsize=32768,timeo=600,retrans=2,intr,addr=149.
>>>>>> 165.226.130)
>>>>>> bl-nas1:/vol/hdln on /N/u type nfs
>>>>>> (rw,nosuid,tcp,rsize=32768,wsize=32768,timeo=600,retrans=2,intr,addr=149.
>>>>>> 165.226.129)
>>>>>> bl-nas2:/vol/soft on /N/soft type nfs
>>>>>> (rw,nosuid,tcp,rsize=32768,wsize=32768,timeo=600,retrans=2,intr,addr=149.
>>>>>> 165.226.130)
>>>>>> bl-nas1:/vol/logs on /N/logs type nfs
>>>>>> (rw,nosuid,tcp,rsize=32768,wsize=32768,timeo=600,retrans=2,intr,addr=149.
>>>>>> 165.226.129)
>>>>>> none on /dev/cpuset type cpuset (rw)
>>>>>> dc-mds01.uits.indiana.edu:/dc on /N/dc type lustre (rw,localflock)
>>>>>> 149.165.235.173:/mds-wan/client on /N/dcwan type lustre (rw,localflock)
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Daniel S. Standage
>>>>>> Ph.D. Candidate
>>>>>> Bioinformatics and Computational Biology Program
>>>>>> Department of Genetics, Development, and Cell Biology
>>>>>> Iowa State University
>>>>>>
>>>>>>
>>>>>> On Fri, May 24, 2013 at 1:29 PM, Carson Holt <carsonhh at gmail.com> wrote:
>>>>>>> They load fine for me. It is an SQLite database. I know that SQLlite
>>>>>>> can freeze on NFS if it's not configured properly.
>>>>>>>
>>>>>>> Could you send me the output from these 3 commands.
>>>>>>>
>>>>>>> ls -al <maker_output directory>
>>>>>>> df <maker_output directory>
>>>>>>> mount
>>>>>>>
>>>>>>> --Carson
>>>>>>>
>>>>>>>
>>>>>>> From: Daniel Standage <daniel.standage at gmail.com>
>>>>>>> Date: Friday, 24 May, 2013 1:13 PM
>>>>>>>
>>>>>>> To: Carson Holt <carsonhh at gmail.com>
>>>>>>> Subject: Re: [maker-devel] Using maker with precomputed transcript /
>>>>>>> protein alignments
>>>>>>>
>>>>>>> I deleted the entire output directory before relaunching. No .db files
>>>>>>> are even created, only the mpi_blastdb directory with the genomic
>>>>>>> sequence data and corresponding index, before it hangs.
>>>>>>>
>>>>>>> The GFF3 files are attached.
>>>>>>>
>>>>>>> Thanks.
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Daniel S. Standage
>>>>>>> Ph.D. Candidate
>>>>>>> Bioinformatics and Computational Biology Program
>>>>>>> Department of Genetics, Development, and Cell Biology
>>>>>>> Iowa State University
>>>>>>>
>>>>>>>
>>>>>>> On Fri, May 24, 2013 at 12:57 PM, Carson Holt <carsonhh at gmail.com>
>>>>>>> wrote:
>>>>>>> Did you delete any *.db files in the maker.output directory first. If
>>>>>>> not do that, and check on the rerun if that file is growing in size. It
>>>>>>> is a database to hold the GFF3 file entries. It's final size should be
>>>>>>> ~ 2x the size of the combined GFF3 files. If it is growing, then it is
>>>>>>> not really frozen (you just need to give it more time). If it is not
>>>>>>> growing, send me your GFF3 files and I can try and duplicate the error.
>>>>>>>
>>>>>>> --Carson
>>>>>>>
>>>>>>>
>>>>>>> From: Daniel Standage <daniel.standage at gmail.com>
>>>>>>> Date: Friday, 24 May, 2013 12:50 PM
>>>>>>>
>>>>>>> To: Carson Holt <carsonhh at gmail.com>
>>>>>>> Subject: Re: [maker-devel] Using maker with precomputed transcript /
>>>>>>> protein alignments
>>>>>>>
>>>>>>> I installed BioPerl-1.6.901, rebuilt Maker, and re-launched the job.
>>>>>>> After running for 10-15 minutes, it seems to be hanging in the same
>>>>>>> place as before.
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Daniel S. Standage
>>>>>>> Ph.D. Candidate
>>>>>>> Bioinformatics and Computational Biology Program
>>>>>>> Department of Genetics, Development, and Cell Biology
>>>>>>> Iowa State University
>>>>>>>
>>>>>>>
>>>>>>> On Fri, May 24, 2013 at 11:38 AM, Carson Holt <carsonhh at gmail.com>
>>>>>>> wrote:
>>>>>>> That is the CPAN version and the last stable release on bioperl.org
>>>>>>> <http://bioperl.org> . Older version as well as the bio-perl live
>>>>>>> version will cause MAKER to fail. The both have issues with the Fasta
>>>>>>> indexing module that maker uses.
>>>>>>>
>>>>>>> http://search.cpan.org/CPAN/authors/id/C/CJ/CJFIELDS/BioPerl-1.6.901.tar
>>>>>>> .gz
>>>>>>>
>>>>>>> --Carson
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> From: Daniel Standage <daniel.standage at gmail.com>
>>>>>>> Date: Friday, 24 May, 2013 11:34 AM
>>>>>>> To: Carson Holt <carsonhh at gmail.com>
>>>>>>> Subject: Re: [maker-devel] Using maker with precomputed transcript /
>>>>>>> protein alignments
>>>>>>>
>>>>>>> I'm not sure if a rebuild of Maker was necessary, but I tried running it
>>>>>>> just to be safe. It's complaining about Bio::Root::Version dependency
>>>>>>> not being met. Looking at the Build.PL file, it requires
>>>>>>> Bio::Root::Version version 1.006901. Is there really such a version, or
>>>>>>> should this be changed to 1.006 or 1.006001?
>>>>>>>
>>>>>>> For now I'll change it to 1.006001 (the installed version) and proceed
>>>>>>> with another test.
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Daniel S. Standage
>>>>>>> Ph.D. Candidate
>>>>>>> Bioinformatics and Computational Biology Program
>>>>>>> Department of Genetics, Development, and Cell Biology
>>>>>>> Iowa State University
>>>>>>>
>>>>>>>
>>>>>>> On Fri, May 24, 2013 at 9:45 AM, Carson Holt <carsonhh at gmail.com> wrote:
>>>>>>> Could you run this command in the maker devel base directory.
>>>>>>>
>>>>>>> svn switch --relocate svn://*
>>>>>>> <http://malachite.genetics.utah.edu/maker/trunk> ************
>>>>>>> svn://* <http://topaz.genetics.utah.edu/maker/trunk> ***************
>>>>>>>
>>>>>>> Then do 'svn update', and then tell me what happens. Make sure to
>>>>>>> delete the and *.db files in the *.maker.output/ directory before
>>>>>>> retrying.
>>>>>>>
>>>>>>> --Carson
>>>>>>>
>>>>>>>
>>>>>>> From: Daniel Standage <daniel.standage at gmail.com>
>>>>>>> Date: Friday, 24 May, 2013 9:10 AM
>>>>>>> To: Maker Mailing List <maker-devel at yandell-lab.org>
>>>>>>> Subject: [maker-devel] Using maker with precomputed transcript /
>>>>>>> protein alignments
>>>>>>>
>>>>>>> Greetings!
>>>>>>>
>>>>>>> I have some precomputed transcript and protein alignments that I would
>>>>>>> like to use with Maker. I have converted them into GFF3 format (see
>>>>>>> attached examples) and provided them to their corresponding entries
>>>>>>> (est_gff, altest_gff, protein_gff) in the maker_opts.ctl file.
>>>>>>>
>>>>>>> Unfortunately, Maker seems to be getting caught up on processing these
>>>>>>> GFF3 files. I've tried running Maker 2.10 as well as the development
>>>>>>> version (checked out a few months ago--svn server isn't responding so I
>>>>>>> can't give a precise revision number), and in both cases Maker hangs
>>>>>>> while trying to create the GFF3 database. These are the last lines I see
>>>>>>> in STDERR when --debug is set.
>>>>>>>
>>>>>>> STATUS: Setting up database for any GFF3 input...
>>>>>>> Calling GFFDB::new at /N/u/dstandag/Mason/local/src/maker-dev/bin/maker
>>>>>>> line 587.
>>>>>>>
>>>>>>> I can't find any documentation specifying any explicit requirements for
>>>>>>> the alignment-containing GFF3 input files. Maker output uses the pretty
>>>>>>> canonical expressed_sequence_match, protein_match, and match_part
>>>>>>> features for encoding alignments, and I have used this convention with
>>>>>>> my input (see attached examples). I have also double-checked that my
>>>>>>> examples are valid GFF3, so my guess is that Maker has additional
>>>>>>> constraints/expectations for certain fields in the GFF3 files (score
>>>>>>> column? required attributes?). Is this correct, and if so would you be
>>>>>>> able to point me toward any related documentation I may have missed?
>>>>>>>
>>>>>>> Many thanks.
>>>>>>>
>>>>>>> --
>>>>>>> Daniel S. Standage
>>>>>>> Ph.D. Candidate
>>>>>>> Bioinformatics and Computational Biology Program
>>>>>>> Department of Genetics, Development, and Cell Biology
>>>>>>> Iowa State University
>>>>>>> _______________________________________________ maker-devel mailing list
>>>>>>> maker-devel at box290.bluehost.comhttp://box290.bluehost.com/mailman/listin
>>>>>>> fo/maker-devel_yandell-lab.org
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://yandell-lab.org/pipermail/maker-devel_yandell-lab.org/attachments/20130524/f774e0ba/attachment-0003.html>
More information about the maker-devel
mailing list