Fig. 6

Pattern spatial sampling and identification of clonal mutations. We compared our model (blue line) to simulation results using random single-cell sampling (red points) and to simulation results using sampling in specific pattern (blue points). Sampling pattern in panel a). Each red point represents the proportion of correct estimates with n=4 samples after 10 000 sampling repetitions. For the green points, the template pattern was rotated to obtain multiple sampling repetitions from each tumour, while maintaining the distance between the cells and the peripheral location of cells. The proportion of correct estimations was calculated from a number of possible sampling repetitions for each tumour. Sampling in pattern appears superior to random sampling throughout most of the range of proportion of largest sub-clone f (Fig. b))