Why tumor samples of a same TCGA donor get different VAF via WGS and ATAC seq?

Question

[solo7773]

This question arose when I read the following science paper

The chromatin accessibility landscape of primary human cancers, 10.1126/science.aav1898

As shown in the figure below (A and E), the tumor samples from a single donor sequenced using WGS and ATAC-seq display different variant allele frequency. From their paper, I did not find out whether WGS and ATAC-seq sequenced the tumor samples of the same TCGA sample UUID. (1) If the sample UUIDs is the same, how can WGS and ATAC-seq get different VAF. If this is possible, it's technical error. This does not convey any biological meaning; (2) If sample UUIDs are different, it's just sampling bias, without any biological sense neither.

Am I wrong? Could anybody help me?

Answer 1

[ATpoint]

ATAC-seq measures open chromatin. If a mutation causes a gain or decrease in transcription factor binding that is associated with chromatin accessability, and the event is heterozygous, then the mutated allele will show different accessability than the wt allele, resulting in different read counts and therefore different VAF. In WGS, there should not be any preference for any of the alleles as you start from "naked" genomic DNA rather than enriching for any regions based on chromatin accessability.

[solo7773]

Thanks ATpoint for your great answer.