Fig 3. Evolutionary rates of substitutions and short indels and their associated 95% confidence intervals (CI) estimated with LSD for the CAO.

Substitution rate was estimated as 1.23e-7 [1.00e-7–1.43e-7] substitutions/site/year and the short indel rate was estimated as 5.37e-9 [1e-10–7.37e-9] indels/site/year. The light colors show the distribution of rates for date-randomized data and their associated lower and upper confidence intervals; information for all randomized data sets are in the supporting information (S8 Fig). For substitutions, there is no overlap in rate and confidence interval between real data and date-randomized data, indicating the presence of temporal signal according to the stringent test. For short indels, the evolutionary rate inferred from the real data does not overlap with the rates of the date-randomized data. However, the confidence interval of the evolutionary rate inferred from the real data overlaps with the upper confidence intervals of the date-randomized data, indicating the presence of temporal signal for short indels according to the simple test. Hence, the temporal signal of short indels is weaker than that of substitutions. Due to the limited number of events, there is not sufficient temporal signal to estimate evolutionary rates for short insertions and deletions separately or for long indels.