“MetaPhaser: methylation-based haplotype phasing of human genomes” is the title of the NCM Houston session I watched tonight. Yilei Fu from Rice University was the presenter. They began sharing how long reads provide more information about structural variations (SV) and help with assembly and phasing. Fu explained an example of thiopurine methyltransferase (TMPT) and how phasing is crucial for detecting important variants. There are three methods for haplotype phasing: population-based, trio-based, and now long-read-based. Long-read phasing uses statistical methods. Long-read phasing does not solve all challenges, however. Homozygosity (identical on two alleles) hinders phasing. MethPhaser improved phase block length by producing SNV phasing results and combining methylation data. They tested MethPhaser on a human genome reference. MethPhaser also improved the detection of medically relevant regions. Haplotype-specific methylation and insertions, for example, are detected. Fu explained that they tested MethPhaser on other samples and populations. They improved N50 length and phase blocks. While this is not my field, I find methylation and phasing intriguing and want to try methylation detection.
