Fig. 1

Overview. A tumor is composed of multiple subpopulations of cells, or clones, with distinct somatic mutations, which can be measured using DNA sequencing. a Due to limitations in inference algorithms and/or sequencing technologies, we are limited to characterizing tumor clones in terms of either single-nucleotide variants (SNVs, stars) or copy-number aberrations (CNAs, triangles). That is, we infer clones \(\Pi _1\), proportions \(U_1\) and a clone tree \(T_1\) for the SNVs. Similarly, we infer clones \(\Pi _2\), proportions \(U_2\) and a clone tree \(T_2\) for the CNAs. b PACTION solves the Parsimonious Clone Tree Integration problem of inferring clones \(\Pi \subseteq \Pi _1 \times \Pi _2\), a clone tree T and proportions U that characterize the clones of the tumor in terms of both SNVs and CNAs