Juhyun Kim, Benjamin D. Rosen, Sarah E Fumagalli, Kristen L. Kuhn, Amy Long, Jeffrey J. Schoenebeck, Heather Schwartz, Lan Wu‐Cavener, Aleksey V. Zimin, Douglas R. Cavener, Timothy P. L. Smith, Adam M. Phillippy, Sergey Koren, Arang Rhie
High-quality reference genomes are critical for studying the biology of the genome, but current methods often leave gaps and errors, especially in repetitive regions. These issues arise from challenges in genome graph curation and are not fully resolvable by standard polishing approaches. To address this, we developed Verkko-Fillet, a Python-based interactive framework for inspecting, editing, and refining genome assembly graphs. It integrates multiple data sources and provides tools for visualization, gap filling, and structural correction. Applied to a giraffe and the benchmark human genome, Verkko-Fillet improves a draft assembly (Q61.5) to a complete telomere-to-telomere genome (Q73.6), increasing both contiguity and accuracy. This work highlights the importance of graph-based curation for producing a finished, gapless genome assembly suitable for downstream analyses.