search

griffithlab / pVACtools

http://www.pvactools.org
BSD 3-Clause Clear License
141 stars 59 forks source link

"Tier" value in the final report is not consistent with "RNA_expr " #727

Closed gudeqing closed 2 years ago

gudeqing commented 3 years ago

Hi:

Describe the bug "Tier" value in the final report is not consistent with "RNA_expr ". I guess this maybe caused by No RNA_VAF provided since you may calculate this by "RNA-expr*RNA_VAF" ?

To Reproduce vcf having gene expression but no RNA_VAF and RNA_Depth

Output File image

Expected behavior "Tier" value in the final report is consistent with "RNA_expr ".

susannasiebert commented 3 years ago

You are correct that the aggregate report's definition of "Expression" is defined as "RNA-expr*RNA_VAF" so if either of those or 0 or NA the variant gets binned in the NoExpr tier. The naming is confusing. We've started to call this new metric "Allele Expression" and with our next release will be including this metric in the aggregate report as an additional output column. Would it makes things easier to understand if we then also renamed the tier to be "NoAlleleExpr", "LowAlleleExpr", etc?

gudeqing commented 3 years ago

@susannasiebert I do not quite understand why we need to define expression as "RNA-expr*RNA_VAF" since this expression is quantified from tumor sample. If No RNA_VAF is available, why not directly set RNA_VAF = 1.0

chrisamiller commented 3 years ago

If you have a cancer point mutation, it's generally going to be heterozygous, affecting only one of the two alleles. Many transcripts undergo allele-specific transcription, where only one of the two alleles is actually expressed. If the WT allele is being expressed (even at very high levels!), but the MT allele is silenced, then it's unlikely to be a good neoantigen candidate. Hope this explanation makes sense!