MGA: a tool for haplotype-mixed assembly of long and accurate reads

Recent large-scale genome sequencing projects have generated near-complete diploid assemblies that reconstruct both haplomes. Producing such assemblies remains formidable, typically requiring large teams, extensive manual curation, and integration of multiple sequencing technologies. However, for many species and applications, a near-complete haplotype-mixed assembly-representing a mosaic of both haplomes-provides most of the same benefits for downstream analyses. Such assemblies can be generated automatically at lower cost using only HiFi reads. Here we present Mosaic Genome Assembler (MGA), a tool that generates near-complete haplotype-mixed assemblies from HiFi reads alone. We show that MGA substantially outperforms existing haplotype-mixed assemblers.