PROJECT ABSTRACT Circulating cell-free DNA (cfDNA) in the bloodstream originates from dying cells and is a promising non- invasive biomarker for cell death. Recent studies have demonstrated that cfDNA levels are correlated with cancer, tissue trauma, autoimmune disease, and organ transplants, indicating the potential clinical utility of cfDNA. CfDNA, however, can originate from any number of tissues throughout the body, and some of the cfDNA present in an individual at a given moment is likely originating from healthy cell turnover. To better understand tissue degeneration, especially in the context of disease, a reliable estimate of the tissue of origin of cfDNA is needed. Identifying the tissue of origin of cfDNA will have direct impact on disease diagnosis and monitoring, and in quantitative biomarker discovery. Pre-existing methods for tissue of origin decomposition are inadequate for cfDNA. Firstly, cDNA is only present in a small amount in the blood. Current decomposition methods generally rely on array-based platforms that require an onerous amount of patient blood, which may not be clinically applicable. An alternative is whole genome bisulfite sequencing data, which requires lower input cfDNA, but is noisy. This data is not addressed with previous methods. Finally, accurately decomposing cfDNA mixtures requires a robust understanding of all possible tissue types that could potentially contribute to the mixture. This robust reference, however, is nearly impossible to assemble, as there are hundreds of distinct tissue types, and because the methylation state for a CpG in a tissue can vary. This could lead to biases in the decomposition results of previous methods. In this proposal, we aim to address the limitations of previous methods by developing a comprehensive workflow for cfDNA tissue-of-origin prediction. We will develop a statistical method for predicting the tissue of origin of cfDNA, allowing for low read count data and unknown tissue types (Aim 1). We hypothesize that this will reduce bias in decomposition estimates and allow our method to be more widely used than previous methods. Additionally, we propose developing a capture protocol that will enrich for cfDNA fragments that are informative of tissue or disease status. This protocol will be designed to use with only small amounts of input cfDNA (Aim 2). We hypothesize that a capture panel approach will vastly reduce sequencing costs and, thus, increase the clinical utility of our approach. Lastly, we propose applying our method to a large cohort of ALS patients and age matched controls (Aim 3). If successful, this biomarker will have substantial impact on the treatment of and drug development for ALS and other degenerative diseases.