TY - JOUR TI - Unsupervised detection of fragment length signatures of circulating tumor DNA using non-negative matrix factorization AU - Renaud, Gabriel AU - Nørgaard, Maibritt AU - Lindberg, Johan AU - Grönberg, Henrik AU - De Laere, Bram AU - Jensen, Jørgen Bjerggaard AU - Borre, Michael AU - Andersen, Claus Lindbjerg AU - Sørensen, Karina Dalsgaard AU - Maretty, Lasse AU - Besenbacher, Søren A2 - Weigel, Detlef A2 - Thierry, Alain VL - 11 PY - 2022 DA - 2022/07/27 SP - e71569 C1 - eLife 2022;11:e71569 DO - 10.7554/eLife.71569 UR - https://doi.org/10.7554/eLife.71569 AB - Sequencing of cell-free DNA (cfDNA) is currently being used to detect cancer by searching both for mutational and non-mutational alterations. Recent work has shown that the length distribution of cfDNA fragments from a cancer patient can inform tumor load and type. Here, we propose non-negative matrix factorization (NMF) of fragment length distributions as a novel and completely unsupervised method for studying fragment length patterns in cfDNA. Using shallow whole-genome sequencing (sWGS) of cfDNA from a cohort of patients with metastatic castration-resistant prostate cancer (mCRPC), we demonstrate how NMF accurately infers the true tumor fragment length distribution as an NMF component - and that the sample weights of this component correlate with ctDNA levels (r=0.75). We further demonstrate how using several NMF components enables accurate cancer detection on data from various early stage cancers (AUC = 0.96). Finally, we show that NMF, when applied across genomic regions, can be used to discover fragment length signatures associated with open chromatin. KW - cell-free DNA KW - cancer genomics KW - liquid biopsy JF - eLife SN - 2050-084X PB - eLife Sciences Publications, Ltd ER -