Fig. 2: Cryptic mtDNA mutations evolve in a clock-like manner consonant with theory and enable inference of mtDNA mutation rate.
a We present the 95% credible interval for each donor’s mitochondrial age as a proportion of the expected time to fixation, and show the maximum a posteriori (MAP) estimate for regression, along with the 95% credible interval of the median inferred coalescent age (see Supplementary Discussion S2). b The posterior on the inferred mutation rate per base per replication is shown under the assumption of 1000 mtDNA per cell, with the MAP estimate highlighted as 4.6 × 10−8. c The proportion of heteroplasmic mutations above a certain threshold reaches equilibrium at the same timescale as human life when modelled using the MAP parameter estimates. d The number of homoplasmic mutations (normalised by bases observed) per cell increases with age in humans (Spearman correlation r≈0.89 and p < 10−7). Also shown is a line generated with the MAP parameters found from fitting the data to our model (Supplementary Discussion S2). We see a similar accumulation of homoplasmic mutations in mice in samples from 2 tissues and 16 mice (Spearman correlation r≈0.85 and p < 10−4). e We look at the first and second derivatives of the MAP estimate for number of homoplasmic mutations against time. These can be roughly equated to the speed and acceleration of ageing. The peak in acceleration of ageing occurs at around 40 years old. f Using the MAP estimate we look at the proportion of cells carrying a mutation above a certain threshold. After around 20 years cells begin to have mutations above a 60% heteroplasmy threshold and by age 40 an appreciable fraction of cells are carrying a mutation with heteroplasmy above 60% and cells have begun to accumulate mutations at homoplasmy (100% heteroplasmy). By the age of 80 nearly 30% of cells are predicted to have a homoplasmic mutations and nearly half are predicted to have mutations at a heteroplasmy above 60%.
