Benchmarking of bioinformatics tools for the hybrid assembly of human and non-human whole-genome sequencing data.

Accurate and complete genome assemblies enable variant identification and the discovery of novel genomic features and biological functions. However, assemblies of large and complex genomes remain challenging. Long-read sequencing data, alone or combined with short-read data, facilitate genome assembly. However, the literature has limited comprehensive evaluations of software performance, especially for human genome assembly. We benchmarked 11 pipelines, including four long-read only assemblers and three hybrid assemblers, combined with four polishing schemes, using the HG002 human reference material sequenced with Oxford Nanopore Technologies and Illumina. The best-performing pipeline was validated with non-reference human and non-human routine laboratory samples. Software performance was assessed using QUAST, BUSCO, and Merqury metrics, alongside computational cost analyses. We found that Flye outperformed all assemblers, particularly with Ratatosk error-corrected long-reads. Polishing improved the assembly accuracy and continuity, with two rounds of Racon and Pilon yielding the best results. The assembly of data from validation samples showed comparable assembly metrics to those of the reference material. Based on the results, a complete optimal analysis pipeline for the assembly, polishing, and contig curation developed on Nextflow is provided to enable efficient parallelization and built-in dependency management to further advance the generation of high-quality and chromosome-level assemblies.