Table 2.

Similarity of the predicted endogenous mitochondrial genome sequence to the original Neanderthal reference sequence, at various rates of simulated contamination with present-day human DNA. An endogenous consensus call was performed using schmutzi on all fragments, and using PMDtools followed by htslib on the fragments labeled by PMDtools as endogenous. For comparison, we generated a simple consensus by running htslib on all sequenced fragments. While this approach works well at low amounts of contamination, it produces an incorrect consensus at higher levels of contamination when the presence of contaminating fragments is not accounted for using approaches like PMDtools and schmutzi. The number of indels are reported as either insertions or deletions in either the predicted consensus or the Neanderthal reference; hence, discrepancies in the final sum may occur

Contamination	Endogenous prediction			Endogenous prediction from			Mitochondrial consensus, called
rate	from schmutzi			PMDtools and htslib			using htslib on all fragments
	Matches	Mismatches	Indels	Matches	Mismatches	Indels	Matches	Mismatches	Indels
1 %	16,565	0	0	16,561	2	6	16,561	3	5
5 %	16,565	0	0	16,561	2	6	16,561	3	5
10 %	16,565	0	0	16,561	2	6	16,561	3	5
15 %	16,565	0	0	16,560	3	6	16,553	11	5
20 %	16,565	0	0	16,560	3	6	16,488	76	5
25 %	16,565	0	0	16,558	5	6	16,374	190	5
30 %	16,564	1	0	16,558	5	6	16,371	193	5
35 %	16,564	1	0	16,556	7	6	16,371	193	5
40 %	16,564	1	0	16,555	8	6	16,371	193	5
45 %	16,564	1	0	16,553	10	6	16,371	193	5
50 %	16,563	2	0	16,553	10	6	16,371	193	5
55 %	16,564	1	0	16,554	9	6	16,370	194	5
60 %	16,563	2	0	16,551	12	6	16,368	196	5
65 %	16,563	1	1	16,551	12	6	16,361	203	5
70 %	16,562	1	2	16,548	15	6	16,358	206	5
75 %	16,563	1	1	16,546	17	6	16,355	209	5
80 %	16,561	2	2	16,545	18	6	16,355	209	5
85 %	16,563	1	1	16,544	19	6	16,355	209	5
90 %	16,561	3	1	16,539	24	6	16,355	209	5
95 %	16,550	15	7	16,532	31	6	16,355	209	5