Fig. 7
Predicting the changing probability of crossing heteroplasmy thresholds over time. Background colour map gives the predicted probability, based on a training subset of data, with which a given heteroplasmy threshold is crossed with time (horizontal axis), for a given initial heteroplasmy (vertical axis). For example, a colour map value of 0.4 at t = 150 dpc, h0 = 0.5 means that 40% of cells in a sample with initial heteroplasmy 0.5 are expected to have crossed the threshold after 150 days. Each blue circle corresponds to a sample of oocytes from a separate test subset of our LE and HB datasets, positioned by mother’s age and reference heteroplasmy. The radius of red circles within these points gives the proportion of oocytes in that sample that cross the threshold. Agreement between prediction and observation is manifest as agreement between the radius of the red circles and the predicted colour map probabilities. (Top) P(h < 0.2), modelling the probability of obtaining a desired low-heteroplasmy (<20%) embryo from PGD; (bottom) P(h > 0.6), modelling the probability with which cells exceed a (60%) threshold for mitochondrial disease
