Extended Data Fig. 4: Examples of authentication for microbial hits.

a, Observed breadth of genomic coverage as a function of average read depth. Coloured symbols indicate hits in species Toxoplasma gondii (left panel) and Yersinia pestis (right panel), with symbol colour indicating relative entropy of read start positions. Solid black line shows theoretical expected breadth of coverage for a given average read depth⁶⁶. b, Lengths of contigs in the reference genome of Toxoplasma gondii and number of samples showing n ≥ 20 reads mapped. Symbol colour indicates the average number of reads mapped to a specific contig across samples. c, Bayesian estimator of aDNA damage (D max) and significance (Z-score) obtained from metaDMG, for hits in species Clostridium botulinum (left) and Yersinia pestis (right). Error bars indicate ± 1 standard deviation, and symbol fill colour indicates average read depth for mapped reads. Samples used as examples in aDNA damage curves (d) are labelled and indicated with black circles. d, aDNA damage patterns for four example hits in species Clostridium botulinum and Yersinia pestis. Plots show observed nucleotide misincorporation frequencies (red symbols and line) and metaDMG fit (black line) and 68% credible intervals (shaded region) for C > T transitions as a function of distance from the 5’ read end.