Suggested pipeline to perform variant calling for cancer exome sequencing samples of PDX models

Question

[svlachavas]

Dear Community,

I would like to ask a methodological question regarding a specific scenario for somatic variant calling concerning whole exome sequencing data:

briefly, based on a biological protocol regarding the isolation of circulating tumor cells (CTCs) from patients with lung cancer, in vivo xenograft experiments were performed- CTCs isolated from specific patients were isolated and injected to mice, and approximately after 6 months, xenografts were removed, cut in small pieces, frozen and gDNA was extracted, which was subjected to WES. At the same time, peripheral blood cell samples (normal) are available from the same patients.

My main question is if there is a suggested pipeline that can be utilized to perform somatic variant calling on these PDX WES data ? using also as reference matched samples the normal PBMC samples from the patients ? I'm aware that one challenge is how to discriminate between mouse and human reads, as also any other infiltration due to stromal infiltration from mouse cells...

Thank you in advance,

Efstathios

Answer 1

[akshayb04]

Hi Efstathios,

This workflow or methodology steps should clear all your doubts.

https://cgc.sbgenomics.com/public/apps#pdxnet/pdx-wf-commit2/pdx-wes-tumor-normal-xenome-with-variant-calling-cnv-estimation-/

In order to split mouse (host) genes from Human (graft) genome, the above tutorial uses Xenome but XenofilteR (https://github.com/PeeperLab/XenofilteR) is a very good R-package to incorporate both exome and rna-seq mapped files simultaneously.

Hope it helps.

Stay safe during the quarantine.

Akshay