In adeschen/aics: Accurate Inference of Genetic Ancestry from Cancer Sequencing

BiocStyle::markdown()

suppressPackageStartupMessages({
  library(knitr)
  library(aics)
})

set.seed(121444)

Licensing

TODO

Citing

If you use this package for a publication, we would ask you to cite the following:

TODO

Installation

To install the latest version accessible on the aics Github Website, the r CRANpkg("devtools") package is required.

## Load required package
library(devtools)

## Install the latest version of aics
devtools::install_github('adeschen/aics')

It is also possible to install an official release. The list of available releases is posted on the aics Release Website.

## Load required package
library(devtools)

## Install the version v0.1.4 of aics
## using 'ref' parameter
devtools::install_github('adeschen/aics', ref = "v0.1.4")

Introduction

Epidemiological evidence point to race and/or ethnicity as important determinants of incidence and outcome in multiple types of cancer [@Cronin2018; @Siegel2020]. Ancestry annotation of cancer data often draws from patient’s self-identified race and/or ethnicity (SIRE). Yet, SIRE information is frequently missing or incomplete [@Nugent2019]. Moreover, SIRE information is not always consistent with genetic ancestry.

Accurate ancestral characterization can also be obtained by genotyping a patient's DNA from a cancer-free tissue. Multiple methods have been implemented to infer ancestry from germline DNA sequence [@Price2006; @Pritchard2000; @Alexander2009]. However, genotyping of DNA from matched normal specimens is not part of standard clinical practice and is not performed routinely outside academic clinical centers. In sum, matched germline DNA sequence is often missing for cancer-derived molecular data. In such cases, having the possibility to infer ancestry from tumor-derived data would be beneficial. However, accurate inference of ancestry is challenging for tumor-derived sequence. Tumor genome are frequently filled with somatic alterations sush as single nucleotide variants (SNV), copy number variants, translocations and microsatellite instabilities.

The aics package implements an inference procedure that has been specifically developed to accurately infer genetic ancestry from cancer-derived sequencing. The covered cancer-derived sequencing are, more specifically, tumor exomes, targeted gene panels and RNA sequences.

The aics package also implements simulation technique that generates synthetic cancer-derived data. Synthetic cancer-derived data can be used to assess the performance of cancer related inference tools.