Fig. 3
From: How many samples are needed to infer truly clonal mutations from heterogenous tumours?
Effect of sample size on clonality inference in well-mixed tumour. The proportion of correct clonality estimates for n=2 samples, each containing 1% of total number of cancer cells. ε represents the frequency bellow which we discard mutations from the clonality analysis. The value of ε where identification of clonal mutations becomes impossible corresponds to the proportion of the largest sub-clone f, as each sample is representative of the whole tumour