Accurate inference of genetic ancestry from cancer-derived data

INTRODUCTION: Genetic ancestry-oriented cancer research requires the ability to perform accurate genetic ancestry inference from existing cancer-derived data, very often in the absence of matching cancer-free genomic data. Here we examine the feasibility and accuracy of such computation. In order to optimize and assess the performance of the ancestry inference for any given input cancer-derived molecular profile, we have developed a data synthesis framework. METHODS: In its core procedure, the ancestral background of the profiled patient is replaced with one of any number of individuals with known ancestry. Data synthesis is applicable to multiple profiling platforms and makes it possible to assess the performance of inference specifically for a given molecular profile, and separately for each continental-level ancestry. This ability extends to all ancestries, including those without statistically sufficient representation in the existing cancer data. IMPLEMENTATION: The inference procedure is implemented as an R package called Robust Ancestry Inference using Data Synthesis (RAIDS). The package uses the genomic data structures provided by the Bioconductor GenomicRanges package and the CoreArray Genomic Data Structure (GDS) data files provided by the Bioconductor SNPRelate, and gdsfmt packages. The data synthesis is part of the package. RESULTS: The inference procedure is accurate and robust in a wide range of sequencing depths. Using three representative cancer types across three molecular profiling modalities, we demonstrate that global, continental-level ancestry of the patient can be inferred with high accuracy, as quantified by its agreement with the golden standard of the ancestry derived from matching cancer-free molecular data. CONCLUSION: Our study demonstrates that vast amounts of existing cancer-derived molecular data potentially are amenable to ancestry-oriented studies of the disease, without recourse to matching cancer-free genomes or patients' self-identification by ancestry. The computational tools for this purpose are now available as an open-source software. Its integration into the Bioconductor is in process. The RAIDS package is available at: https://github.com/KrasnitzLab/RAIDS