Distinguishing how genetics impact cellular processes can improve our understanding of variable risk for diseases. Although single-cell omics have provided molecular characterization of cell types and states on diverse tissue samples, their genetic ancestry and effects on cellular molecular traits are largely understudied. Here, we developed Monopogen, a computational tool enabling researchers to detect single nucleotide variants (SNVs) from a variety of single cell transcriptomic and epigenomic sequencing data. It leverages linkage disequilibrium from external reference panels to identify germline SNVs from sparse sequencing data and uses Monovar to identify novel SNVs at cluster (or cell type) levels. Monopogen can identify 100K~3M germline SNVs from various single cell sequencing platforms (scRNA-seq, snRNA-seq, snATAC-seq etc), with genotyping accuracy higher than 95%, when compared against matched whole genome sequencing data. We applied Monopogen on human retina, normal breast and Asian immune diversity atlases, showing that that derived genotypes enable accurate global and local ancestry inference and identification of admixed samples from ancestrally diverse donors. In addition, we applied Monopogen on ~4M cells from 65 human heart left ventricle single cell samples and identified novel variants associated with cardiomyocyte metabolic levels and epigenomic programs. In summary, Monopogen provides a novel computational framework that brings together population genetics and single cell omics to uncover genetic determinants of cellular quantitative traits.