Documentation error in RNAseq workflow

Posted by Geraldine_VdAuwera on 11 Jun 2014 (11)

We discovered today that we made an error in the documentation article that describes the RNAseq Best Practices workflow. The error is not critical but is likely to cause an increased rate of False Positive calls in your dataset.

The error was made in the description of the "Split & Trim" pre-processing step. We originally wrote that you need to reassign mapping qualities to 60 using the ReassignMappingQuality read filter. However, this causes all MAPQs in the file to be reassigned to 60, whereas what you want to do is reassign MAPQs only for good alignments which STAR identifies with MAPQ 255. This is done with a different read filter, called ReassignOneMappingQuality. The correct command is therefore:

java -jar GenomeAnalysisTK.jar -T SplitNCigarReads -R ref.fasta -I dedupped.bam -o split.bam -rf ReassignOneMappingQuality -RMQF 255 -RMQT 60 -U ALLOW_N_CIGAR_READS

In our hands we see a bump in the rate of FP calls from 4% to 8% when the wrong filter is used. We don't see any significant amount of false negatives (lost true positives) with the bad command, although we do see a few more true positives show up in the results of the bad command. So basically the effect is to excessively increase sensitivity, at the expense of specificity, because poorly mapped reads are taken into account with a "good" mapping quality, where they would normally be discarded.

This effect will be stronger in datasets with lower overall quality, so your results may vary. Let us know if you observe any really dramatic effects, but we don't expect that to happen.

To be clear, we do recommend re-processing your data if you can, but if that is not an option, keep in mind how this affects the rate of false positive discovery in your data.

We apologize for this error (which has now been corrected in the documentation) and for the inconvenience it may cause you.

sirian

Thanks for the correction! I was actually wondering a little bit why you changed every score.

sboyle

Thanks for the correction Geraldine!

Geraldine_VdAuwera

@kam This may be a good recommendation to make to the authors of the mappers.

justinjj

Dear Geraldine, Could you please clarify me is there any difference or issue if I use the mapq as 50 instead of 60 as suggested to run GATK? The bwa aligner higher quality is 60 but the tophat2 (uses bowtie2) provide higher quality/unique mapq as 50 and GATK runs without any error in both cases also by star aligner "--outSAMmapqUnique 50" The RNAseq mappers is already giving meaningful quality score? https://software.broadinstitute.org/gatk/guide/article?id=3891 (So we use the GATK’s ReassignOneMappingQuality read filter to reassign all good alignments to the default value of 60. This is not ideal, and we hope that in the future RNAseq mappers will emit meaningful quality scores, but in the meantime this is the best we can do.) Thanks.

Geraldine_VdAuwera

Yes, that's fine. If newer versions of the mappers produce MAPQ scores in that range, there is no need to reassign a different value.

justinjj

Thanks Geraldine.

NeillGibson

Hi, How can I do the re assignment of the STAR mapping quality in GATK4? I have some trouble finding the correct documentation / an GATK4 SplitNCigarReads CLI example. See also https://gatkforums.broadinstitute.org/gatk/discussion/10800/gatk4-how-to-reassign-star-mapping-quality-from-255-to-60-with-splitncigarreads/ Thank you.

Geraldine_VdAuwera

@NeillGibson Since this was published many moons ago, the STAR aligner gained the ability to assign meaningful mapping quality scores on request. So we did not implement the ability to reassign MAPQs into GATK4.

NeillGibson

@Geraldine_VdAuwera Thank you for this information. I passed this information on to the issue that I started at bcbio, where I found out first about this issue. https://github.com/chapmanb/bcbio-nextgen/issues/2163 One risk that I see is that using the STAR --outSAMmapqUnique 60 option maybe fixes the issue with GATK, but that other downstream tools maybe still depend on the (still default) STAR mapping quality value of 255 (e.g. cufflinks). > The mapping quality MAPQ (column 5) is 255 for uniquely mapping reads, and int(-10*log10(1- > 1/Nmap)) for multi-mapping reads. This scheme is same as the one used by TopHat and is com- > patible with Cuffinks. The default MAPQ=255 for the unique mappers maybe changed with > --outSAMmapqUnique parameter (integer 0 to 255) to ensure compatibility with downstream tools > such as GATK. > https://github.com/alexdobin/STAR/blob/master/doc/STARmanual.pdf

Geraldine_VdAuwera

Ah, that’s fair concern. I do think that eventually we will want to port the read transformer that made it possible to change the mapqs, but to be frank it’s not a high priority for us. You can still use the read transformer capability using PrintReads in the old version as a stopgap in the meantime.

Version highlights for GATK version 3.1...

- Upcoming events

See Events calendar for full list and dates

- Recent events

See Events calendar for full list and dates

GATK Dev Team

@gatk_dev

Some really great points here. We teach multiple workshops every year; while the short format is not the best possi… https://t.co/r2EoyKNBFK
14 Dec 17
@nickschurch @pathogenomenick @nanopore Heeey... well you may have a point. We don't really do dirty ;) Never tried… https://t.co/zTH5QI89XZ
14 Dec 17
What's new in GATK4? In this short video, Laura Gauthier explains how the speed and scalability of joint calling is… https://t.co/EAPdrZcqJI
14 Dec 17
What is truth? Or, how an accident of nature can illuminate our path https://t.co/qperntHUVz
8 Dec 17
Why is HaplotypeCaller slower in the most recent GATK4 beta versions? Because it's saving its strength for the 4.0… https://t.co/p3ZBPojxkR

- Our favorite tweets from others

My first @gatk_dev blog-post https://t.co/JwVzAtqDvt about the CHM cell-lines experiment we now use as benchmark da… https://t.co/8VdqG6H4Zk
9 Dec 17
Wanna be a baller, HaplotypeCaller 20K genotypes in the VCF file Caller, gettin' phased tonight https://t.co/bOGSI4UL23
17 Nov 17
This amazing genomics toolkit helps researchers find insights that save lives - I know! GATK users - please provide… https://t.co/gdY4FDPX8K
2 Nov 17
using GATK to identify SNPs while handing out candy... Happy Halloween! @broadinstitute @gatk_dev #bioinformatics #researchisfun #Halloween
31 Oct 17
Although it made me cry sometimes, I owe them a lot and love them much more. https://t.co/vUj0cBllgn
16 Oct 17
See more of our favorite tweets...