[bam_parse_region] fail to determine the sequence name

[SS00]

I'm trying to run differential expression using EdgeR and/or DEseq2 on the Ratsch Lab galaxy server but I keep getting this error:



[bam_parse_region] fail to determine the sequence name


I mapped the groomed and filtered FastQ files with Tophat2 using the mm10 reference and did DE expression with the Tophat2 bam files and the UCSC genes.gtf files.



Anyone know what could be the problem?

[dpryan]

At what point did it give you that error? There's actually no reason that that function within samtools should ever be used for computing differential expression (at least none that I can think of).

[SS00]

So from the log of EdgeR (but also DESeq2) I can see that it gets up to three steps:

% 1. Data preparation %

% 2. Read counting %

% 3. Differential testing %

and then it gives an error at step 3.

In addition to the [bam_parse_region] error as follows:

[bam_parse_region] fail to determine the sequence name.
Invalid region chrY_random:54420149-54423069
R script execution failed

It also says the following at the end of the log file:

Error in `row.names<-.data.frame`(`*tmp*`, value = value) :
duplicate 'row.names' are not allowed
Calls: rownames<- -> row.names<- -> row.names<-.data.frame
In addition: Warning message:
non-unique values when setting 'row.names': ‘4933409K07Rik’, ‘A430089I19Rik’, ‘AU018829’, ‘AY761185’, ‘BC002163’, ‘BC061212’, ‘Ccl21b’, ‘Ccl21c’, ‘Cngb1’, ‘E330014E10Rik’, ‘Gm10591’, ‘Gm13298’, ‘Gm13304’, ‘Gm13308’, ‘Gm15056’, ‘Gm15085’, ‘Gm15093’, ‘Gm16367’, ‘Gm1993’, ‘Gm3286’, ‘Gm5506’, ‘Gm5512’, ‘Gm5643’, ‘Gm5801’, ‘Gm6040’, ‘Gm6367’, ‘Mir138-2’, ‘Mir1906-1’, ‘Mir1906-2’, ‘Mir684-1’, ‘Obox2’, ‘Ott’, ‘Snord58b’, ‘Sult1c1’, ‘Tagap’, ‘Tff1’, ‘Tsnax’, ‘Ube1y1’, ‘Vmn1r186’, ‘Vmn1r187’, ‘Vmn1r62’, ‘Vmn1r63’
Execution halted

It's driving me crazy.

Thanks for your help!

At what point did it give you that error? There's actually no reason that that function within samtools should ever be used for computing differential expression (at least none that I can think of).

[dpryan]

Did you write that galaxy pipeline or did someone else?

[SS00]

Someone else. I use it on https://galaxy.cbio.mskcc.org/ under Differential/Quantitative Analysis.

I used Tophat2 on the same server, downloaded igenomes UCSC genes.gtf and ran DESeq2 and EdgeR with the generated bam files and the downloaded gtf.

[dpryan]

Short answer: Whoever wrote that pipeline didn't know what they were doing.

Longer answer: It looks like you aligned to a genome that lacked chrY_random, whereas the annotation file has that. That's what leads to the "[bam_parse_region]" error. Having said that, that shouldn't occur because I expect this pipeline is performing the counting incorrectly. This is also what's leading to the error at the end when it's trying to create a dataframe. Basically, don't use that pipeline. You have three options for moving forward. (1) Use a different galaxy pipeline. You already have the BAM files, so this is probably doable. (2) Download the aforementioned BAM files and do things correctly locally (this would require you to know how to analyze your dataset). (3) Collaborate with a local bioinformatician. This is the best idea, since if you're using galaxy you're probably new to this. Galaxy is convenient, but if you're new to things then it's a black box that just spits out results that may, or may not, be correct (and you'd likely have no way of knowing which).

[SS00]

Thank you! That makes sense.

Yeah, I really have no idea of analysis so I agree my best bet is to work with one of our bioinformaticians and get it sorted.