Miranda Galey from the University of Washington spoke at London Calling 2023 about “MeOW: genome-wide identification of differentially methylated regions using Oxford Nanopore long-read sequencing data.” Galey explained how long-read sequencing can shorten the time to diagnosis. The noted that methylated nucleotides have different current patterns that can be detected. The advantage of long-read sequencing is that variants, methylation, and phasing can be evaluated. Galey explained that “CpG islands are stretches of DNA with CG:GC ratio >0.6 often observed near promotor regions.” The need to quickly identify methylation differences lead Galey to develop the Methylation Optimization Wizard – MeOW: a command-line tool that works in conjunction with R. The sequencing depth needed should be >30x. Input bam files need to have methylated regions encoded using HTS5 standards. Intervals of sequences should be annotated. In R, an “enormous amount” of comparisons are performed. MeOW calculates paired or pooled variance t-tests. Beta test statistics are used to improve sensitivity. Cohen’s D, and graphs are plotted. Galey shared two examples from cases of Prader-Willi (PWS) and Angelman syndrome (AS). They used a PromethION flow cell and evaluated samples with CpG islands at 100 regions. MeOW was used with unknown cases, and emphasized that visual inspection of the results is needed for confirmation. Galey wants to integrate phasing into MeOW and validate with other samples. I think tools like this one will help increase use of methylation analysis for diagnostics, as they make interpreting the data less daunting!
