Peptide Ancestry Informative Markers in Uterine Neoplasms from Women of European, African and Asian Ancestry

Abstract Deep proteogenomic analyses of diverse populations will improve our understanding of molecular drivers of disease, such as those underlying uterine neoplasms which exhibit marked racial disparities. The characterization of ancestry-linked peptide variants in disease-relevant patient tissues represents a foundational step towards investigating relationships linking patient ancestry with the proteome, genome and disease pathogenesis. We selected 1,037 nonsynonymous single nucleotide polymorphisms (nsSNPs) encoding missense amino acid substitutions occurring within putative tryptic peptides exhibiting high allele frequencies (≥ 50%) in European, African, and East Asian populations, termed peptide ancestry informative markers (pAIMs) and investigated these in cell line models and tissues of uterine neoplasms collected from women representing diverse ancestries. We quantified 62 pAIMs in total across >100 patient-derived samples by multiplexed, quantitative proteomic analyses. Our analyses of endometrial cancer cell lines established from European (n=7), African (n=3), and East Asian (n=3) women resulted in the quantitation of 43 pAIMs that significantly correlated with population-level allele frequencies consistent with global ancestry estimates defined by genotype analyses. We next investigated uterine leiomyoma (ULM) tissues from self-described African-American (n=31) and European-American (n=35) women and quantified 33 pAIMs that correlated with patient race. We validated 10 of these in an independent cohort of ULM tissues collected from women determined to be of European (n=16) and African (n=11) ancestry by genotype analyses. This validated subset was predominated by candidates of undetermined clinical significance (ClinVar) and included a substitution in GC vitamin D-binding protein, i.e. (D451E, rs7041), that exhibits a higher allele frequency in European (58%) versus African (7.3%) populations and correlates with altered GC protein and vitamin D levels. Comparison of cell line and tissue analyses revealed co-quantitation of pAIMs encoded within Protein-L-isoaspartate(D-aspartate) O-methyltransferase (V178I, >80% AF in African and East Asians), Serine/threonine-protein phosphatase CPPED1 (A19D, >75% AF in Africans) and AH receptor-interacting protein (Q228K, >99% AF in European and East Asians). We further assessed the performance of pAIMs to determine patient ancestry in silico relative to standard genotype estimates by random sampling analyses and find that as few as 20 population-specific pAIMs can determine ancestry proportions similarly to >260K SNPs (R2=0.9905). These efforts have quantified and validated pAIMs in diverse uterine neoplasms that represent a foundational set of ancestry-linked variant peptides observable by multiplexed proteomics and provides proof of concept that estimates of European, African and East Asian ancestry can be determined from routine proteomic analysis of disease-relevant patient cell lines and tissues. This work will support ongoing efforts focused on linking the proteome and genome with patient ancestry to better understand relationships with disease mechanisms driving racial disparities in the pathogenesis of uterine neoplasms.