Table 1. List of parameters used to describe the inference model in multiSNV.

Parameter	Description	Comments
n	Number of tumour samples	User-specified
SN	Allelic composition in normal sample	Estimated (one to two alleles allowed)
	Allelic composition in tumour sample i	Estimated (one to three alleles allowed)
k	Denotes sample k	k ∈ {N, T1…Tn}
Mk	Number of alleles in allelic composition Sk	Computed from inferred Sk
	Pileup of bases and corresponding base qualities in all samples	Read from pileup file
	Pileup of bases and corresponding base qualities in normal sample	Read from pileup file
	Pileup of bases and corresponding base qualities in tumour sample i	Read from pileup file
	Allele supported by read j in sample k	Read from pileup file
	Error probability of read j in sample k	Read from pileup file
	Total number of reads in sample k that support allele τ	Read from pileup file
	The probability distribution of alleles in sample k	Estimated by maximizing
	Vector of shape parameters of Dirichlet prior on	Uniform prior
μ	Mutation rate	User-specified (default is 3 × 10−7)
	The set of values of SN with nonzero prior probability	All monoallelic and diallelic compositions (total of 10)
ϕz	Sampling probability of allelic composition	Integrated out
δz	Parameters of Dirichlet prior on ϕz of	As described in Materials and Methods
	Parameters of Dirichlet prior on ϕz of SN	As described in Materials and Methods
	Number of tumour samples excluding Ti with allelic composition	From most recent draw of the Gibbs sampler
nz	Number of tumour samples with allelic composition	From most recent draw of the Gibbs sampler
	Scales pseudocounts for Dirichlet prior of allelic compositions	10 × (n + 1)

multiSNV analyses each location in the genome independently, so these parameters refer to a single genomic locus.