What level BAM should Manta be run on

[icanwinwyz]

Hello,

Can I run Manta on a GATK realigned and recalibrated BAM or in the original BWA mapped bam? Would the results vary drastically if we ran it on both?

Thanks!

[riederd]

Hi,

I just noticed that difference is huge. When I use recalibrated and indel realigned tumor/normal BAMs from a WGS sample, I get only 17 somatic SVs (somaticSV.vcf.gz). When I then use the BWA aligned and markduplicate BAM of the same sample I get 1261 SVs.

Can someone comment on this? What result should we trust more?

Perhaps also of interest: We have also matched RNAseq data from that sample, and when we call gene fusions using arriba the "overlap" with the manta WGS result is much better with the manta results produced from BAMs that were not indelrealigned and reaclibrated.

Thanks Dietmar

[riederd]

Oh... I just realized, that the difference is due to only using coding regions for making the realigned BAM file. Sorry.