Abstract 4568: Biochemical mitigation of technical artifacts improves performance of a ctDNA methyl sequencing assay

In addition to accumulating mutations that may drive tumor growth, cancer cells also show broad changes in DNA methylation patterns genomewide. For example, a colorectal cancer sample may have fewer than 10 driver mutations, fewer than 100 passenger mutations, but may have nearly 800 differentially methylated regions. In contrast to the traditional method of using sequencing to find these scarce mutations in the genome, probing for these relatively abundant aberrantly methylated regions in the genome can be a powerful alternative approach for the detection of cancer. However, assessing DNA methylation via methyl sequencing is complicated by artifacts introduced during the traditional NGS library prep process. In vivo, CpG sites in mammals are typically methylated on both strands of the double stranded DNA; however, we observe that the majority of read fragments display some level of hemimethylation (only one strand displays methyl modifications), and roughly 10% of all reads are hemimethylated across the read. Furthermore, we observe that the fraction of methylated bases gradually decreases over the course of a sequencing read; the methylation state of DNA should not be expected to be impacted by the cycle number of a sequencing run. These artifacts are collectively known as M-bias and can be mitigated computationally, but these methods are an incomplete solution and may sacrifice valuable base coverage. We have devised a strategy to correct M-bias biochemically, using a custom formulated end repair mix. We show that this novel library prep workflow reduces M-bias by ∼50% in diverse ctDNA assays. This reduction in M-bias results in more accurate methylation status calls and reduces the number of reads needed to confidently cover a CpG region of interest. Finally, we show that biochemical M-bias mitigation methods lower healthy hypomethylation background signal by ∼10x. These changes increase robustness of hypomethylation biomarker quantitation and utility in ctDNA detection assays. Citation Format: James Yu, Peter George, Jason Close, Andy Eang, Neil Peterman, Yu Sun, Nafisa Bulsara, Eileen Dimalanta, Alexander Robertson, Yexun Wang, Nicole Lambert. Biochemical mitigation of technical artifacts improves performance of a ctDNA methyl sequencing assay [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4568.