Fig. 6

Fitted deviations in pseudogene differences for high-coverage assemblies. Reads corresponding to very high coverage (≥ 1000-fold) assemblies were reassembled with Unicycler to explore the effect of taxonomy and coverage on the generation of spurious pseudogenes. Reassemblies differed considerably in the density of spurious pseudogenes at 50-fold coverage (top row) and the coverage required to reach only a single pseudogene difference per Mbp (bottom row). This result implies considerable variability in spurious pseudogene counts in real data beyond that observed for simulated reads, which are shown by the leftmost points. Only species with at least 13 independent read sets (n) are shown, with the remainder merged into the “Other” category (rightmost points)