RT Journal Article
SR Electronic
T1 R<span class="sc">epair</span>S<span class="sc">ig</span>: Deconvolution of DNA damage and repair contributions to the mutational landscape of cancer
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 2020.11.21.392878
DO 10.1101/2020.11.21.392878
A1 Damian Wojtowicz
A1 Jan Hoinka
A1 Bayarbaatar Amgalan
A1 Yoo-Ah Kim
A1 Teresa M. Przytycka
YR 2020
UL http://biorxiv.org/content/early/2020/11/22/2020.11.21.392878.abstract
AB Many mutagenic processes leave characteristic imprints on cancer genomes known as mutational signatures. These signatures have been of recent interest regarding their applicability in studying processes shaping the mutational landscape of cancer. In particular, pinpointing the presence of altered DNA repair pathways can have important therapeutic implications. However, mutational signatures of DNA repair deficiencies are often hard to infer. This challenge emerges as a result of deficient DNA repair processes acting by modifying the outcome of other mutagens. Thus, they exhibit non-additive effects that are not depicted by the current paradigm for modeling mutational processes as independent signatures. To close this gap, we present RepairSig, a method that accounts for interactions between DNA damage and repair and is able to uncover unbiased signatures of deficient DNA repair processes. In particular, RepairSig was able to replace three MMR deficiency signatures previously proposed to be active in breast cancer, with just one signature strikingly similar to the experimentally derived signature. As the first method to model interactions between mutagenic processes, RepairSig is an important step towards biologically more realistic modeling of mutational processes in cancer. The source code for RepairSig is publicly available at https://github.com/ncbi/RepairSig.Competing Interest StatementThe authors have declared no competing interest.