ATaRVa: Analysis of Tandem Repeat Variation from Long Read Sequencing data

Tandem Repeats (TRs) are contiguous repetitions of DNA motifs whose variations modulate key cellular processes and traits. Their expansions are causally linked to over 70 neurological disorders in humans. Historically, genotyping TRs has been challenging due to their low sequence complexity and allele lengths exceeding typical short-read sequencing capabilities. While long-read sequencing improves TR genotyping by spanning large expansions and sequencing native DNA molecules, existing long-read genotypers often are inaccurate, computationally inefficient, and show platform-specific biases. Here, we present ATaRVa, a genotyper that achieves superior accuracy across both PacBio and Oxford Nanopore platforms, while running [~]15x faster than existing tools. Beyond genotyping, ATaRVa features sequence-level motif decomposition and methylation profiling. Evaluated across diverse datasets and genome-wide catalogs, ATaRVa especially excels in low coverage datasets, and accurately identifies known pathogenic expansions in clinical samples. Importantly, by quantifying uninterrupted target motifs in a population scale analysis, the tool systematically resolves false positives caused by benign, interrupted alleles. In addition, we present VisuaMiTRa, an auxiliary interface that enables interactive visualization of motif-level decomposition and base-level methylation of the TR alleles. By combining high computational efficiency with sequence-level structural awareness, ATaRVa provides a highly scalable framework for TR analysis in both population cohorts and clinical settings.