SHAKIRA: Scaled hierarchical ancestry K-population inference from RNA-seq alignments

Background Genetic ancestry estimation is an underutilized analytical approach that provides valuable biological context for studies investigating population-associated variation in gene expression, disease biology, and treatment response. Admixture-based ancestry estimates have historically relied on DNA microarray data, whole-genome sequencing, and whole-exome sequencing. However, RNA sequencing (RNA-seq) data contains informative markers suitable for broad ancestry estimates. Given the widespread availability of RNA-seq datasets, a standardized and accessible RNA-seq-based ancestry estimation pipeline would expand opportunities to incorporate genetic ancestry into transcriptomic analyses. Results Herein, we describe Superpopulation-level Haplotype Ancestry from K-population Inference of RNA Assays (SHAKIRA), a bioinformatic pipeline for estimating global genetic ancestry from cancer-derived RNA-seq data and evaluate its performance using both experimental tumor RNA-seq samples and an external reference dataset. First, we applied the pipeline to RNA-seq data from 86 pancreatic neuroendocrine tumor samples and compared predominant ancestry coefficients to self-reported race. Using a 60% admixture proportion threshold to assign predominant ancestral superpopulation, 38 of 39 patients who self-reported as Black were assigned predominant African ancestry, and all 47 patients who self-reported as White were assigned predominant European ancestry. We then externally validated the pipeline using 144 Cancer Cell Line Encyclopedia (CCLE) samples with published DNA microarray–based ancestry estimates. Ancestry estimates derived from RNA-seq data from CCLE samples demonstrated strong concordance with DNA-based estimates for African, American, and European ancestry components, with Pearson correlations of 0.995, 0.979, and 0.994, respectively. Conclusions These findings support the feasibility of using RNA-seq-derived variants to estimate broad genetic ancestry, particularly for African and European superpopulation components. SHAKIRA provides a practical approach for incorporating genetic ancestry into RNA-seq-based studies when matched DNA sequencing or genotyping data are unavailable.