Fig. 1

Schematic of germline imputation from low-coverage sequencing. a The unobserved germline variants to be imputed, with 0/1/2 reflecting the number of ALT (non-reference) alleles. b The sequenced input data with targeted regions (dark blue), on-target reads aligning to these regions (light blue), and off-target reads typically discarded (orange). c The haplotype reference panel as a matrix of 0/1 alleles (corresponding to reference or ALT). Alleles that match sequenced reads are shown in light orange, alleles that form a haplotype match shown in dark orange, and alleles that reside along the haplotype but did not carry reads in the target sample shown in gray. d The imputed germline variant dosages/probabilities, quantifying the expected number of non-reference alleles. In sum, the matched haplotype is used to refine the dark orange alleles and impute the light gray alleles in the target individual