VariantEval - explanation of output fields?

Jeremy37

Hello, I've just started using VariantEval. It seems very useful and provides a good deal of output. However, it's not obvious to me what all of the fields mean. Do you know when there will be documentation of the output?

For example, some things I'd like to know:

VariantSummary:

• What is the difference between TiTvRatio and TiTvRatioPerSample?
• What is SNPNoveltyRate? (In general I don't understand the rate fields)

TiTvVariantEvaluator:

• What is TiTvRatioStandard? What are the nTiDerived and nTvDerived fields?

CountVariants:

• What are variantRate and variantRatePerBp? Is variantRatePerBp the average number of bp between variants? (If anything it seems like the names for these should be reversed.)
• What are heterozygosity and heterozygosityPerBp?

What is the --dbsnp parameter used for? Is this redundant with --comp?

Many thanks for your help,

Jeremy

Jeremy37

Also, what is the format for the --stratIntervals argument? I searched for IntervalBinding[Feature], but can't find an explanation.

I might like to specify different allele frequency bins for the AlleleFrequency stratification, e.g. 0.10 - 0.15, rather than bins in increments of 0.005 all the way from 0 to 1.0.

Geraldine_VdAuwera

Hi Jeremy,

Unfortunately our docs are a little thin on this topic right now; we're going to try to plump them up in the near future. In the meantime, the best way to get a description of the output is to look at the comments in the source code for each stratification module. You can view them at this link:

https://github.com/broadgsa/gatk/tree/master/public/java/src/org/broadinstitute/sting/gatk/walkers/varianteval/stratifications