This technology is new method for determining tissue-of-origin of ctDNA using a novel assay to detect methylation patterns. Bisulfite sequencing is the standard method for detecting methylation patterns; however, it cannot be used to accurately measure ctDNA because of degradation issues with very low levels of input DNA. A recent study discovered that methylation status significantly influences DNA fragmentation patterns, and these inventors took advantage of this observation to generate a fragment-length algorithm to identify tissue-of-origin. To generate the computational model, the algorithm was trained to recognize methylation-dependent fragmentation patterns using whole genome bisulfite sequencing of DNA with a known methylation pattern. The algorithm can be used to identify tissue-of-origin of ctDNA by analyzing the fragmentation pattern of sequencing reads in the sample then using the fragment lengths to predict the methylation pattern of the ctDNA. Additionally, this technique can be used with ultra-low-pass sequencing of 0.1x, which significantly reduces sequencing costs.