Login / Signup

Scalable telomere-to-telomere assembly for diploid and polyploid genomes with double graph.

Haoyu ChengMobin AsriJulian LucasSergey KorenHeng Li
Published in: Nature methods (2024)
Despite advances in long-read sequencing technologies, constructing a near telomere-to-telomere assembly is still computationally demanding. Here we present hifiasm (UL), an efficient de novo assembly algorithm combining multiple sequencing technologies to scale up population-wide near telomere-to-telomere assemblies. Applied to 22 human and two plant genomes, our algorithm produces better diploid assemblies at a cost of an order of magnitude lower than existing methods, and it also works with polyploid genomes.
Keyphrases
  • machine learning
  • endothelial cells
  • deep learning
  • convolutional neural network
  • herpes simplex virus