Inferring clonal heterogeneity in cancer using SNP arrays and whole genome sequencing

Bioinformatics (2019) doi: 10.1093/bioinformatics/btz057

In the original article, under heading 2.4 ‘Performance metrics’, the equations under points 2, 3 and 4 were incorrect. These have now been corrected as below.

• 2. The distribution (as a fraction of the tumor) of cancer cells among the existent clones; the distance metric between true and estimated $\psi$ (j being the index of the clone) is:

  $$D_{\psi }^{}=\sqrt{\sum _{j=1}^K{({\psi }_j-{\psi }_{\mathit{j,true}})}^2}$$
• 3. The detection of CNVs in each clone; the weighted distance between the true and estimated copy numbers for clone j in segment i:

  $$D_{\mathit{CN}}=\sqrt{\sum _{i=1}^N\sum _{j=1}^K{w}_{\mathit{ij}}{\left({\mathit{CN}}_{\mathit{ij}}–{\mathit{CN}}_{\mathit{ij},\mathit{true}}\right)}^2}$$

  The weight terms, w_ij, are defined as:

  $${\mathrm{w}}_{\text{ij}}={ \mathrm{N}}_{\mathrm{i}}* {\Psi }_{\mathrm{j}}$$

  where N_i = the length of the ith segment (so an erroneous copy number inference is weighted more severely for longer segments).
• 4. The detection of mutations in each clone. The distance metric indicating the distance between the true and estimated number of mutated copies at each locus of a putative somatic mutations, again weighted according to the size of the clone is as follows:

  $$D_M=\sqrt{\sum _{i=1}^N\sum _{j=1}^K{w}_{\mathit{ij}}{\left(M_{\mathit{ij}}–M_{\mathit{ij},\mathit{true}}\right)}^2}$$

  where i denotes the mutation and j denotes the clone. The weights w_ij are defined as for copy numbers above. However, in the case of mutations, the segment length N_i = 1 for all mutations.