Plasma ChIP-seq cell-free nucleosomes Based Identification Of Genomic Amplification And Their Driver Genes

Genomic amplifications are one of the hallmarks of cancer growth. Copy number variations (CNV) in tumor DNA are highly informative about disease progression and often dictate treatment choices. Methods developed in the past decade demonstrate the utility of circulating tumor DNA (ctDNA) to estimate tumor CNV directly from plasma samples, sparing the need for invasive tumor biopsy. However, such analysis does not pinpoint the genes active in the CNV and thus can be uninformative about driver genes. Here we show how to use a new liquid biopsy method to answer both questions from one assay. We recently introduced a method of chromatin immunoprecipitation and sequencing of cell-free DNA (cfChIP-seq) to predict gene transcription in the dying cells contributing to circulation. Here we present a new statistical model to evaluate genomic amplification from cfChIP-seq data, which is informative on both tumor CNV and transcription in the dying tumor cells. We characterize the link between the presence of an increased copy number, that is, an amplification, to the active driver genes in the amplification region. Analyzing cohorts of colorectal cancer and small cell lung cancer patients, we identify known CNVs (e.g., MYC and ERBB2 amplifications) and many less understood amplifications with their driver genes. In contrast, we did not identify cancer-related amplifications in self-declared healthy donors. With this method, it is possible to identify genes that are crucial in the development of the tumor and to track changes in the CNV and activated genes during the patient treatment in a non-invasive manner.