Copy number variants.

The table has one row per CNV, with coordinates, width and type (gain, loss) shown. Each CNV is assigned a basic ID and an extended ID. The basic numeric ID is unique to the pair of coordinates and type (gain, loss), regardless of phylogenetic relationship and may be present in multiple rows (with the exception of partial back mutations, which share the same CNV ID as the parent CNV, but whose coordinates are within the parent CNV footprint). The extended ID is unique per row and incorporates additional information with the following suffixes: R (recurrence, a repeated phylogenetically unrelated CNV involving the same coordinates and type; each recurrence is numbered), BM (back mutation, a CNV which returns a phylogenetically earlier CNV to CN2 or another lower amplitude state), PBM (partial back mutation, similar to BM, but the CNV coordinates lie within the parent CNV footprint), G (gain, higher amplitude gain beyond CN3; each independent gain is labelled, i.e., G1, G2, etc., and the copy number state is also labelled, i.e., G1CN4), and L (similar to G, but for losses below CN1). The samples and clade groups of samples carrying each CNV are listed in the “Sample_group” column, together with the copy number state assigned. Subclonal copy number states are expressed as a 0.5 step between the two closest integer states. “Copy_number” lists each state found in samples carrying the CNV, and the number of samples with copy number at each state are shown. In cases where back mutations have occurred, final copy number state is presented also in the parent CNV, including in cases where the back mutation is partial. CNVs that occur uniquely in cell lines or in cell lines experimentally inoculated into animals are tagged, and “Cancer_type” specifies if the CNV occurs in DFT1, DFT2, or non-DFTD tumours. The “subclonal” tag marks CNVs that have non-integer copy number; however, these may also include integer copy number states in tetraploid tumours (tagged). CNVs that occur exclusively within tetraploid tumours are marked, together with the whole genome duplication (WGD) event number in parentheses (S4 Fig); CNVs that are inferred to have occurred in a common ancestor of a tetraploid tumour group prior to WGD are marked; others are inferred to have occurred after WGD. Linkage groups tag groups of CNVs that cooccur in the same direction (gain, loss), and CNVs that involve the same coordinates but the opposite direction (gain, loss) are marked. Nested CNVs are CNVs that have arisen within the coordinates of a “parent” CNV and are in the same direction (gain, loss) as the parent. The total number of tumours carrying each CNV is listed in the “Number_of_tumours” column. CNVs that were part of the M5 complex or which were part of high-level amplicons were not used to generate the phylogenetic tree (Fig 1 and S1 Fig) and are tagged in the “Do_not_use_to_build_tree” column. We have tagged intervals that include physically map markers (labelled “Deakin_2012_markers” and “Taylor_2017_markers”) [17,31]. The number of tumours carrying each CNV as a fraction of the total number of tumours within each clade group is shown, as well as individual genotypes for all tumours in the set (“1,” presence; “0,” absence). Copy number plots labelled with CNV IDs can be found in S2 Data in https://doi.org/10.5281/zenodo.4046235. (XLSX)