Billy Lau from Stanford University School of Medicine spoke at the Nanopore Community Meeting in Houston about “Nanopore sequencing of cell-free DNA for methylation-based breast cancer detection.” I have watched Lau speak at other ONT events. This session was about a research study to improve cancer detection. Lau spoke about breast cancer detection issues. Liquid biopsies offer a non-invasive cancer measurement. You can examine proteins, DNA, and even circulating cells. Lau and team focused on DNA methylation. Tumor suppressor genes have epigenetic reprogramming changes. In general, Lau explained that there are thirty million CpG sites in the human genome that can be analyzed. Bisulfite treatment and sequencing have been extensively used; however, these approaches cause DNA damage, GC skew, artifacts from conversion, and complicated analysis. With Nanopore sequencing, methylation patterns can be obtained by applying models to analyze raw data. Small amounts of cell-free DNA (cfDNA) can be sequenced. Lau shared data from patients and half a milliliter of plasma. Cancer patient-derived cfDNA is enriched in mononucleosomes. After conducting a proof-of-concept pilot, Lau expanded to a larger study and a case-control cohort with 800+ patients. This resulted in over 50 Tb of raw sequencing data that was processed with Stanford’s HPC cluster. Super-accuracy base-calling models with Dorado server were produced. To produce a classifier, the team used a train/test scheme with the “leave-one-out validation.” The classification procedure is a random forest. The classifier performance was analyzed to determine the sensitivity/specificity threshold. The results are encouraging, and Lau is examining selected biomarkers. They are validating the approach on new cohorts. In addition, Lau and team are improving performance by adding additional omic data. This was another intriguing and inspiring Lau session.
