Merging multiple bam files to perform genome-guided assembly

[Lucila Traverso]

Hi all,

I am new to Trinity and trying to perform a genome-guided assembly. The Trinity version I am using is v2.1.1. I have 8 PE samples. As I first aligned my reads with STAR, I have one bam file for each sample, so 8 in total. I ran samtools merge to obtain a single bam file. My question is: Do I have to run samtools sort after this, or the resulting file of the merge step is enough to use in --genome_guided_bam?

Once I have this done, I am planning to run Trinity with all the 16 fq files and the merged bam to obtain a single assembly. Is my approach correct?

Thanks a lot for your help and advices.

Best,
Lucila.

[Brian Haas]

Hi Lucila,

After you merge the bam files, you'd need to coordinate-sort them.  (samtools sort)

Before going through all this, just be sure to read through:

https://github.com/trinityrnaseq/trinityrnaseq/wiki/Genome-Guided-Trinity-Transcriptome-Assembly

You might want to also compare running Stringtie and doing full genome-free denovo assembly to see how they all fare.

best,

~brian

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrnaseq-users@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

--

--
Brian J. Haas
The Broad Institute
http://broadinstitute.org/~bhaas

 

[Lucila Traverso]

Thank you so much, Brian!
I have another question: I am using paired end reads for the assembly, but also some single end reads (that survived the Trimmomatic step as unpaired). I joined them all in a single file (with cat) and added them as --left. But I haven't aligned them with STAR, so the bam file only contains the coordinates for the paired end-reads. Do you think that it can be a problem? Would it be better to also align them and add that information to the bam file?

Thank you again, this group is very helpful for me!

Lucila.

 

[Brian Haas]

Hi Lucila,

I haven't tried mixing paired and single-end data w/ the genome guided pipeline.   I'm not anticipating any trouble, but wouldn't be surprised if it broke.  You can give it a try.   You'd need to align the PE and SE reads separately, then merge and coord-sort before using it as input.

If it breaks, I could probably make it work w/ a little effort.

best,

~b

--

You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrnaseq-users@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

[Lucila Traverso]

I have run it as I told you, with paired and single end reads, but only with the paired reads bam file. The phase 1 was completed succesfully, but the phase 2 gave me an error that appears a lot of times (42MB log file) and starts with:

** The inchworm process failed.sh: /usr/local/bioinformatic/trinityrnaseq-2.1.1/Inchworm/bin//inchworm: No such file or directory

then it finishes saying


We are sorry, commands in file: [FailedCommands] failed.  :-(

Error, cmd: /usr/local/bioinformatic/trinityrnaseq-2.1.1/trinity-plugins/parafly/bin/ParaFly -c trinity_GG.cmds -CPU 10 -v  died with ret 256 at /usr/local/bioinformatic/trinityrnaseq-2.1.1/Trinity line 2183.

I think that this is an installation problem so I am trying to solve it. Then I will try with your advice.

Thank you so much,
Lucila.

[Brian Haas]

yeah, looks like an installation issue.

Running 'make' 

or 'make clean && make'

in the base installation directory should solve this.

best,

~b

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrnaseq-users@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

[Lucila Traverso]

Thanks Brian, it worked! Apparently with the mix of paired- and single-end input files.
But I have a doubt: after Phase 2, the program stopped and the log file ends with:

All commands completed successfully. :-)

Friday, April 13, 2018: 11:59:30    CMD: find Dir_*  -name '*inity.fasta'  | /usr/local/bioinformatic/trinityrnaseq-2.1.1/util/support_scripts/GG_partitioned_trinity_aggregator.pl TRINITY_GG > Trinity-GG.fasta.tmp

Finished. See Trinity-GG.fasta for reconstructed transcripts


I was expecting Phase 3 to start. Is something wrong with my files? Is something missing in my assembly results?

Thank you so much.
Lucila.

[Brian Haas]

Hi Lucila,

It's all good.  There's just 2 phases in Trinity v2:   (1) partitioning reads into clusters, and (2) assembling the reads.   But there are 3 primary programs involved, which is where Trinity gets its name from.

best,

~b

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrnaseq-users@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.