Problem when run TGICL. Is tgicl not able to work with some certain sequence quantities of fasta file ?

Question

[zengxi.hada]

I was running TGICL using fasta file to do sequence cluster and assembly. I encountered some ridiculous problems. It is that tgicl did work in some certain sequence quantities, but cannot work for some other certain sequence quantities. I don't know why. I will describe the problem in detail at following.

I ran tgicl with a fasta file with 239635 sequences.

$ tgicl -F Trinity_all_in_one_3.fasta
Use of :locked is deprecated at /usr/share/perl5/TGI/DBDrv.pm line 36.
tgicl (Trinity_all_in_one_3.fasta) encountered an error at step ASSEMBLE
                 Working directory was /media/zengxi/ffc280cc-f4ce-455b-888d-43dcbb2ff4c3/ubuntu/project/sanqi/filter_contig/all_in_one.

When I extracted the 10000 header line of the Trinity_all_in_one_3.fasta, it worked!

$ head -10000 Trinity_all_in_one_3.fasta > Trinity_all_in_one_3.fasta.head10000
$ tgicl -F Trinity_all_in_one_3.fasta.head10000 
Use of :locked is deprecated at /usr/share/perl5/TGI/DBDrv.pm line 36.
tgicl (Trinity_all_in_one_3.fasta.head10000) finished on machine 
                 in /media/zengxi/ffc280cc-f4ce-455b-888d-43dcbb2ff4c3/ubuntu/project/sanqi/filter_contig/all_in_one, without a detectable error.

However, when I extracted the 100 header line of Trinity_all_in_one_3.fasta, I didn't work. Sounds ridiculous.

$ head -100 Trinity_all_in_one_3.fasta > Trinity_all_in_one_3.fasta.head100
$ tgicl -F Trinity_all_in_one_3.fasta.head100
Use of :locked is deprecated at /usr/share/perl5/TGI/DBDrv.pm line 36.
tgicl (Trinity_all_in_one_3.fasta.head100) encountered an error at step ASSEMBLE
                 Working directory was /media/zengxi/ffc280cc-f4ce-455b-888d-43dcbb2ff4c3/ubuntu/project/sanqi/filter_contig/all_in_one.

All above mentioned input files are in correct fasta format.

I also try another two fasta files with 215933 and 215933 sequences respectively. They all work well using tgicl.

I don't know why. Did Anyone encounter this kind of case before? How can I make the Trinity_all_in_one_3.fasta with 239635 sequences work well using tgicl? I really appreciate it if anyone can help me to solve the problem.

Answer 1

[jorvis]

I'm definitely battling similar issues. I started by replacing the version of cap3 it was distributed with, and that fixed some assemblies but not others. One assembly even completed when I re-ran with increased cores, which also didn't make any sense. Still looking into this, and will post again if I make any headway.

[zengxi.hada]

Hi, some people else also encounter this problem like us. It seems like a program bug.

[meetliuyating]

Did you solve the 'error at step ASSEMBLE' successfully? I finish clustering step,but meet same error as you at following.

$ tgicl -F trinity_all_with_bowtie.fasta
 clustering [trinity_all_with_bowtie.fasta] finished at Jul 30 08:55:07 2017
           The clusters are stored in file 'trinity_all_with_bowtie.fasta_cl_clusters'.

ASSEMBLE [trinity_all_with_bowtie.fasta] started at Jul 30 08:55:30 2017
           WAITING for all children to finish before starting last child!
           WAITING for all children to finish!

 Process terminated with an error, at step 'ASSEMBLE'!
tgicl (trinity_all_with_bowtie.fasta) encountered an error at step ASSEMBLE
                 Working directory was /home/liuyt/rna-seq/tgicl.

And since the cap3 is installed with Tgicl,I have no idea how to update it solely.How did you achieve it? By the way,when the first time error with ASSEMBLE，there is an newest version of cap3 in my miniconda.

[meetliuyating]

Oh.I updated my CAP3,and it assembled successfully.