Fig. 4.

ROH estimates for the 0.5× imputed and the high-quality genotypes. a) The total length of ROH ≥ 2 Mb and ROH < 2 Mb (includes ≥0.5 Mb) plotted with a subset of the reference panel as a background. Filtering included MAF ≥2.5%, transversions only, and no missingness. b) The total length of ROH split into length bins for the seven test samples. The number of sites used for analysis = 481,786 filtered for a MAF ≥2.5%, transversions only, and no missingness. c) ROH along Chr12 for the four ancient genomes at 0.5×, and high-quality demonstrating the general consistency between the imputed and high-quality data set; filters are the same as b).