Finishing complete genomes from telomere to telomere with Verkko-Fillet

Authors: KIM, JUHYUN - National Institutes Of Health (NIH); KOREN, SERGEY - National Institutes Of Health (NIH); Rosen, Benjamin; Kuhn, Kristen; FUMAGALLI, SARAH - Oak Ridge Institute For Science And Education (ORISE); ZIMIN, ALEKSEY - Johns Hopkins University; SCHOENEBECK, JEFFREY - National Institutes Of Health (NIH); WU-CAVENER, LAN - Pennsylvania State University; CAVENER, DOUGLAS - Pennsylvania State University; Smith, Timothy; PHILLIPPY, ADAM - National Institutes Of Health (NIH); RHIE, ARANG - National Institutes Of Health (NIH)

Submitted to: Cell Genomics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/20/2026
Publication Date: N/A
Citation: N/A

Interpretive Summary: Modern techniques for livestock improvement rely heavily upon detailed knowledge of an animal's genetic code. We now have the ability to write out the entire genetic instructions for an animal in what is known as a reference genome. Generating these reference genomes is still incredibly complex, and our existing methods often leave out important pieces or have errors, especially in tricky, repetitive parts of the genetic code. This means we aren't getting the full picture we need. We have developed a new software tool, called Verkko-Fillet, that acts like a sophisticated editing suite for these reference genomes. Imagine trying to build a complex puzzle where some pieces are missing or in the wrong place. Our new tool helps scientists easily see, fix, and fill in those missing or incorrect pieces in the genetic map. We've used it to significantly improve the reference genome of a giraffe, taking it from a good draft to a nearly perfect, complete reference. This means we can now create much higher quality, gap-free genetic maps for any animal. This new tool is a game-changer for genetics research. For our agricultural sector, it means we can generate far more accurate genetic maps for livestock and crops. This will accelerate the development of plants and animals that are more resistant to disease, grow more efficiently, and produce better quality products, ultimately saving farmers untold dollars each year and improving food security for consumers. This investment in a fundamental scientific tool will have widespread benefits across agriculture ensuring we get the most valuable information possible from our genetic studies.

Technical Abstract: High-quality genome assemblies are essential for understanding speciation, evolution, and building references and pan-genomes. Despite major advances in long-read sequencing technologies and graph-based assembly algorithms, current assemblies often remain incomplete and require graph path refinement, including correction of haplotype switching and resolution of gaps—particularly within repetitive regions such as ribosomal DNA clusters and recent segmental duplications. These limitations largely stem from challenges in graph curation and are not adequately addressed by conventional polishing methods, which focus on nucleotide-level corrections. To overcome these issues, we developed Verkko-Fillet, a Python-based interactive framework for genome graph inspection, editing, and curation and provides a step-by-step guideline for downstream polishing. Verkko-Fillet uses output files from Verkko, including the assembly graph, haplotype paths, Hi-C contacts, and additional ONT reads or alternative assemblies aligned to the graph. It offers tools for visualizing, modifying, and exporting curated assembly graphs. It enables users to track changes, resolve complex structural features, fill gaps, and enhance assembly quality after polishing. Using the giraffe genome (Giraffa tippelskirchi) as a case study, we demonstrate how to upgrade a draft assembly (QV 61.5) to a telomere-to-telomere reference genome (QV 73.6) by improving contiguity and completeness using Verkko-Fillet. Our results highlight the critical role of a graph-based curation in producing high-quality, gapless telomere-to-telomere assemblies that are ready for downstream analysis.