A computational framework to simultaneously quantify DNA methylation, somatic copy number alternation and DNA heterogeneity from low coverage plasma circulating DNA sequencing

Genome of Hepatocellular Carcinoma (HCC) undergoes profound changes, including DNA hypomethylation and somatic copy number alternations (SCNA). These two characteristics provide orthogonal information for HCC early diagnosis, and can be assessed by whole-genome bisulfite sequencing (WGBS) of the plasma circulating DNA. We proposed a computational framework to simultaneously quantify DNA methylation and SCNA from plasma circulating DNA WGBS, and further estimate the heterogeneity of the circulating DNA. Our approach reliably detected global DNA hypomethylation and SCNA from low coverage WGBS of tumor and plasma circulating DNA from HCC subjects compared to healthy control individuals. The chromosomal pattern of SCNA detected from tumor DNA and plasma DNA are highly consistent. The computational framework we proposed make efficient use of WGBS and able to simultaneously characterize DNA hypomethylation SCNA, which provide orthogonal evidence in HCC early diagnosis. Importantly, our approach estimated the tumor DNA fraction in plasma circulating DNA, ranging from 38.55% to 1.79%, and is correlated with tumor size (Spearman's correlation coefficient = 0.68, p-value=0.0049). We estimate that the tumor DNA content in plasma could be below 2% for HCC tumor of 2cm or smaller in diameter, which requires relatively high coverage WGBS for reliable assessment.