#3 · Epigenomics

Benchmarks and tools expose technical biases in sequencing-based genomic assays

The common thread is the evaluation and correction of technical bias in sequencing-based measurements. ATAC-seq peak calling is shown to be distorted by local DNA dosage from copy number and replication timing, and the authors propose MACS with WGS control. MethylBench compares six methylation technologies across GIAB and blood and fibroblast samples, while TUSCO uses a curated single-isoform gene set as internal ground truth to benchmark long-read transcriptome reconstruction. What is new is grounding method assessment in endogenous or cross-platform references rather than assuming accuracy. HERRO-corrected Nanopore Simplex reads combined with Verkko2 yield telomere-to-telomere human genome assemblies. STARCall and rhapsodist provide open-source workflows for in situ sequencing and BD Rhapsody single-cell RNA-seq data. These works address adjacent rather than identical problems.