Fig. 3: Global and local ancestry inference using single-cell genotypes derived by Monopogen.

a,b, Genetic ancestry of the four retina samples using Monopogen genotypes derived from snRNA-seq data (a) and genotypes from matched WGS data (b). Colored dots represent individuals in the HGDP reference panel, and black crosses represent the retina samples. The variance explained by PC1 and PC2 from the HGDP panel was labeled. c,d. Local ancestry inference of a European sample 19D013 using genotypes from the snRNA-seq (c) and the WGS (d) data. The 3,202 phased genotypes from 1KG3 were used as the reference for local ancestry inference. Colors in each chromosome denote the inferred source ancestry with a bin size of 1 centimorgan (cM). e,f, Local ancestry results from an admixed sample 19D015. g,h, Local ancestry inference accuracy for 19D013 (g, overall score: 0.96) and 19D015 (h, overall score: 0.90). Each dot denotes the ancestry accuracy for each segment (1 cM). i, PCA-projection analysis shows the ancestry of samples in the AIDA and the HBCA cohorts. j, UMAP of Korean and Japanese samples in the AIDA using genotypes called Monopogen. The UMAP was constructed based on the top five PCs of Korean and Japanese genotypes (on 584,164 SNVs). k, Concordance between Illumina GSAv3 genotyping array data and Monopogen calls across the AIDA samples. Darker colors denote a higher level of concordance between two data modalities. Calculation of the concordance scores is detailed in Methods.