Abstract
Recent advancements in long-read sequencing technologies are renowned for providing extended read lengths and lower error rates, which enhance the assembly of complex genomes. However, high costs and stringent sample quality requirements limit their widespread adoption, especially for degraded DNA samples. In contrast, short-read technologies require shorter DNA fragments but produce reads challenging genome assembly continuity. Reference-guided assembly offers a practical solution by aligning contigs with a reference genome, thereby improving scaffold continuity. However, the reference-guided assembly can introduce more misassemblies. To address this limitation, this study explores using Ragtag's Correct function integrated with in silico libraries to correct misassemblies in reference-guided assemblies. Using three draft genomes from two fish species, we demonstrate that this hybrid strategy significantly improves scaffold assembly accuracy. Specifically, in Megalobrama amblycephala, misassemblies were reduced from 8298 to 4920, and cross-links between different chromosomes decreased from 192 to zero in the corrected assemblies. In two Culter alburnus draft genomes, misassemblies were reduced from 5689 and 6582 to 4728 and 5861, respectively, while cross-links between different chromosomes were significantly reduced from 132 and 13 to five and ten in the corrected assemblies. This approach allowed precise correction of scaffold assembly errors, showcasing its potential to enhance the accuracy of genomic assemblies. Our findings underscore the importance of integrating additional genomic data to achieve reliable genome assemblies, especially for species with significant structural variations. This research provides valuable insights into optimizing genome assembly processes, contributing to advancements in genomic studies.