Table 7.

Downsampling experiment for simulation of low coverage data

Sampling	0.005	0.01	0.02	0.05	0.10	0.20	0.30	0.40
Expected mean coverage	0.10	0.19	0.39	0.97	1.94	3.88	5.82	7.76
Achieved mean coverage	0.09	0.19	0.38	0.94	1.88	3.75	5.63	7.51
Cov % of genome ≥1 read	8.70 %	16.47 %	29.71 %	56.28 %	76.85 %	88.41 %	90.69 %	91.28 %
Cov % of genome by ANGSD	8.64 %	16.37 %	29.55 %	56.06 %	76.68 %	88.35 %	90.67 %	91.27 %
% of correctly called bases	99.72 %	99.73 %	99.74 %	99.77 %	99.82 %	99.89 %	99.92 %	99.93 %

The first row refers to the percentage of reads that were randomly sampled from the original data set (LBK/Stuttgart) from Lazaridis et al. [19] (coverage ≈19×) using SAMtools. The expected mean coverage was derived by multiplication of the original coverage with the sampling value. The achieved mean coverage was calculated using QualiMap after mapping. Cov % of genome ≥1 read is the percentage of the genome that was covered by at least one read. This was also calculated using QualiMap. Cov % of genome by ANGSD is the percentage of the genome that was reconstructed from the genotypes as derived with ANGSD. % of correctly called bases is the percentage of bases that ANGSD called correctly not regarding the base “N”