Split accepted_hits.bam file after Tophat run?

[hong_sunwoo]

Hello.

I am working on two RNA seq data from different conditions.
Now I am using tophat - cufflinks pipe line.

At first time, I ran tophat seperately each sam files (previous version Tophat was used).

Yesterday, I learned that two sample can be pooled in tophat process and ran tophat(v 1.1.0) as below.
$ tophat /rnaseq/bowtie/indexes/hg19 s_1_sequence.txt,s_2_sequence.txt

After run, I found that only one Bam file (accepted_hits.bam).
Because two RNA-seq data was processed, I guess that Tophat might report to Bam files.
Does it mean that I misunderstood something?
Or is it possible to split up again as quote below?

If you do pool the reads, you could also rename them to tag them by sample, so you can split the sample alignments up again after the TopHat run if needed.

[rocksd]

I believe you need to remove the "," between the two sequencing file if they are from two different samples.

[krobison]

Removing the comma between files is not compatible with the manual; supplying two lists like that will mean they will be interpreted as the two sides of paired end reads & if the ids don't match up I suspect the program will error out.

Usage: tophat [options]* <index_base> <reads1_1[,...,readsN_1]> [reads1_2,...readsN_2] 


<reads1_1[,...,readsN_1]>	 A comma-separated list of files containing reads in FASTQ or FASTA format. When running TopHat with paired-end reads, this should be the *_1 ("left") set of files.
<[reads1_2,...readsN_2]>	 A comma-separated list of files containing reads in FASTA or FASTA format. Only used when running TopHat with paired end reads, and contains the *_2 ("right") set of files. The *_2 files MUST appear in the same order as the *_1 files.

[simonandrews]

We've had the same problem. We'd like to process multiple files in the same batch - using the combined evidence from all files to do junction detection. At the moment we've run things like you said, and then parsed the sam output file to use the ids to decide which original file the hit came from. However this won't work in all cases since the ids aren't always unique between different files.

I think adding this functionality would be really useful, and did suggest this to the developers, but haven't heard anything back as yet.

[krobison]

Could you ensure uniqueness of ids by prefixing them prior to running through TopHat? (Not that I love one more preprocessing step)

[fkrueger]

Yes prefixing to ensure uniqueness would work.

[simonandrews]

Could you ensure uniqueness of ids by prefixing them prior to running through TopHat? (Not that I love one more preprocessing step)

You could, but it's a pain because you'd have to duplicate all of your original files to do this. Also, when we've tried this off larger numbers of samples (more than 8 lanes worth) then tophat seems to come to a grinding halt when it's trying to do the junction detection. It seems to create enormous temporary files which it then spends ages trying to sort. We ended up killing it after 24hours at this step.

If this is to work efficiently I suspect there'd need to be some more structural changes inside the program.