Project Summary Non-invasive detection of cell-free DNA(cfDNA) promises to impact clinical regimens of a wide range of diseases, e.g. prenatal conditions, cancer, transplantation, autoimmune disease, trauma, and cardiovascular disease. While this field is emerging as one of the most promising and exciting areas of medicine, few bioinformatics tools are available to facilitate the information extraction from cfDNA sequencing data, although cfDNA data possesses many unique properties. In this proposal, we aim to generate a suite of computational methods facilitating the analysis and interpretation of cfDNA sequencing data, and demonstrate its utilities in cancer detection and characterization. Specifically, we will develop computational methods for the following applocations: (1) Ultra-sensitively detect and locate multiple types of cancer using cfDNA methylome; (2) Detect Copy Number Variation (CNV) in cfDNA sequencing data; (3) Annotate Single Nucleotide Variations (SNV) in cfDNA sequencing data. These computational tools will be validated with cfDNA samples collected from a cohort of lung cancer patients participating in an immunotherapy clinical trial, a repository of blood samples from patients with different types of cancer, and a cohort of liver cancer patients. Although we use cancer as the main context for developing these applications, many of the methods can be adapted to other diseases, e.g. prenatal diagnosis and organ transplant monitoring. We expect that the above open-source tools will significantly facilitate cfDNA- based disease diagnosis and monitoring. 1

Public Health Relevance This project will develop novel computational methods for the analysis and interpretation of cfDNA sequencing data. The methods will significantly facilitate the information extraction from cfDNA data, and push forward the field of cfDNA-based diagnostics and monitoring.