GATK 'Variant quality score recalibration' for shallow coverage

[jorge]

Hello,

I've used GATK to produce variant calls (VCF files). I have shallow coverage (<10) and the documentation suggests using 'Variant quality score recalibration' according to 'true sites' from HapMap 3 and Omni 2.5M SNP chip array. My question is, will I loose all the variants not occuring on those controls? For example, will I loose novel variant?

any advise on this is welcome

Thanks
Jorge

[mamons]

Hello,

those sets with "true sites" are used to build an error model then each snp will be given a score based on the error model and the other parameters that has been collected.

So you will not loose any information.

[jorge]

Hello!

Thanks a lot for clarifying that!

Jorge