Fig. 1

Flowchart illustrating the process of defining a mutation’s status in multiple related samples. Begin with an input data matrix containing the numbers of mutant and wildtype read counts for every mutation in all related samples (top), the mutation's status in each sample is classified in a two-step fashion. First, positive sample/mutation pairs were identified. We assume the users have completed this step using their preferred method before running MSN. The MSN method does not create or remove positive statuses but directly report them to the output (left). Second, for every mutation, each “non-positive” sample is compared with every positive sample to determine if they may contain the same frequency of mutant reads. If and only if this null hypothesis is rejected against all positive samples, then this non-positive sample is considered as negative (right), otherwise, it would be classified as unknown due to low coverage (middle). The output is a data matrix containing all updated mutation statuses (bottom)