search

deweylab / RSEM

RSEM: accurate quantification of gene and isoform expression from RNA-Seq data
http://deweylab.biostat.wisc.edu/rsem/
GNU General Public License v3.0
408 stars 118 forks source link

inconsistent read length #56

Open ESjokvist opened 7 years ago

ESjokvist commented 7 years ago

Hello,

I'm running RSEM-1.3.0, rsem-calculate-expression and getting the error 'Paired-end read K00166:58:H7WLVBBXX:4:1223:31751:40965 has alignments with inconsistent mate lengths!' I'm guessing this originates in read trimming, so am I not supposed to do read trimming before mapping? I've used Skewer for adapter and quality trimming. Should I use a different trimmer or none at all? If I remove read pairs with different lengths I get only 25% of my reads left for that specific sample, which is non-ideal.

Thanks, Elisabet

manueltar commented 6 years ago

Hi,

I am having a similar problem.

I am using Rsem 1.3 and STAR 2.5. My reads are 250 bp and the seq is paired end. I compiled STAR using make STARlong.

I depend on the alignment produced by the alignment pipeline team in the Sanger Institute. They take care of trimming and aligning and I get the results as bam file.

Because I wanted to realign using a gtf file from Ensembl I sorted by read name and then used bamToFastq to convert my sorted bam file into two fq files, each one corresponding to the paired end r1 and r2 files.

I run rsem-calculate-expression as: rsem-calculate-expression \ -p 20 \ --star \ --output-genome-bam \ --calc-ci \ --star-output-genome-bam \ --estimate-rspd \ --paired-end \ $fq1 \ $fq2 \ $reference \ $output

This manages to get me an output with TPMs per transcript but no posterior mean counts or any other parameter calculated by --calc-ci.

Opening the report file I see that the pipeline exited in the rsem-tbam2gbam step: rsem-tbam2gbam: BamConverter.h:131: void BamConverter::process(): Assertion `cqname != qname' failed.

After googling a bit about it I have used rsem-sam-validator on the transcript.bam outputed by Rsem. And I get to the same type of error of ESjokvist:

Read MS3_23843:1:1101:2273:11831 have alignments showing different read/mate lengths! The input file is not valid!

Can you provide input? Thank you very much

sommerJY commented 6 years ago

Did anybody solve the problem??

samadelk commented 1 year ago

these issues have been closed without any answer ! several years later i'm getting the same problem