Broad Institute TCGA MAF procedure

Question

[umn_bist]

I am calling snv and indels on tumor/matched normal RNA-seq data from TCGA using Mutect2.

I am cross-validating with the curated MAF file by Broad, available at TCGA, but I was curious to know what sort of validation/QC/filtering steps they go through as a lot of their variants are not showing up in my raw/filtered vcf file.

Is there any documentation of their pipeline/workflow at TCGA?