Benchmarking Hi-C scaffolders using reference genomes and de novo assemblies.

Area of Science:

• Genomics
• Bioinformatics

Background:

• High-throughput sequencing necessitates high-quality reference genomes for new species studies.
• Chromosome-length genome assembly remains a significant challenge.
• Chromatin conformation capture (Hi-C) methods have advanced genome scaffolding, but lack comprehensive benchmarking.

Purpose of the Study:

• To benchmark the performance of popular Hi-C scaffolding tools.
• To evaluate scaffolder capabilities under ideal and real-world conditions.
• To identify strengths and weaknesses of leading Hi-C scaffolding software.

Main Methods:

• Literature review to identify prominent Hi-C scaffolders (Lachesis, HiRise, 3d-dna, SALSA, AllHiC).
• Performance testing on four reference genomes (S. cerevisiae, L. tarentolae, A. thaliana, H. sapiens).
• Scaffolding of both fragmented reference genomes and de novo long-read assemblies.

Main Results:

• All tested scaffolders achieved >80% accuracy under ideal conditions.
• Performance significantly degraded under more challenging conditions.
• Several scaffolders exhibited poor performance even on near-chromosome-length assemblies.
• HiRise and Lachesis showed the best average performance across all tested scenarios.

Conclusions:

• Hi-C scaffolders exhibit variable performance depending on data quality and assembly stage.
• HiRise and Lachesis are recommended for robust genome scaffolding.
• This study provides crucial insights for selecting appropriate Hi-C scaffolding tools.