Stelka Somatic output

[pinnintil...@gmail.com]

Hi

I'm currently working how to build NGS pipeline based on comparison of aligners vs. callers. I performed analysis using strelka2 on breast cancer dataset normal/tumor (no. of threads - 8) after analysis it generated a vcf for snps and indels both are of same size 9.8 mb. I need confirmation, whether it ran successfully ?

Input -

normal.bam - 63 GB

tumor.bam - 79 GB

hg38.fa - 3.1 GB

Output -

snps.vcf & Indels.vcf - 9.8 mb

CMD -
configureStrelkaSomaticWorkflow.py --normalBam=/data/shayantan/cancer_samples/sample1/SAM/likith/markduplicatesfixed_Realigned.normal.bam --tumorBam=/data/shayantan/cancer_samples/sample1/SAM/likith/markduplicatesfixed_Realigned.tumor.bam --referenceFasta=/data/shayantan/Ref/hg38.fa --outputCallableRegions --runDir=/data/shayantan/cancer_samples/sample1/SAM/likith/StrelkaSomaticWorkflow

StrelkaRunStatsReport -
TotalHours    29.8232h wall, 29.6455h user + 0.0619h system = 29.7075h CPU (99.61%)

CallRegionCandidateIndels    954843
CallRegionNonCandidateIndels    6200436

best,

Likith

IIT Madras

[Saunders, Chris]

Hi Likith,

 

Thanks for checking. Note this email list is inactive and any current strelka issues should be reported through github here:

 

https://github.com/Illumina/strelka/issues

 

Some quick notes on your command below:

 

1. The somatic callable regions bed track is an experimental feature with a very high runtime cost (runtime with this flag is approx. 2x normal runtime), please be sure you really want this feature on. More discussion in the user guide is here:

 

https://github.com/Illumina/strelka/tree/master/docs/userGuide#somatic-callability

 

2. For hg38 we recommend restricting calling to the primary chromosome assemblies per the instructions here:

 

https://github.com/Illumina/strelka/tree/master/docs/userGuide#improving-runtime-for-references-with-many-short-contigs-such-as-grch38

 

….without such masking the runtime can be more variable as the caller occasionally gets stalled in certain high depth decoy regions.

 

Best,

 

-Chris

--
You received this message because you are subscribed to the Google Groups "strelka-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to strelka-discu...@googlegroups.com.
To post to this group, send email to strelka...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/strelka-discuss/8c439b50-e20e-496c-954e-f189d79abfe0%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.