OMACC: an Optical-Map-Assisted Contig Connector for improving de novo genome assembly

BACKGROUND: Genome sequencing and assembly are essential for revealing the secrets of life hidden in genomes. Because of repeats in most genomes, current programs collate sequencing data into a set of assembled sequences, called contigs, instead of a complete genome. Toward completing a genome, optical mapping is powerful in rendering the relative order of contigs on the genome, which is called scaffolding. However, connecting the neighboring contigs with nucleotide sequences requires further efforts. Nagarajian et al. have recently proposed a software module, FINISH, to close the gaps between contigs with other contig sequences after scaffolding contigs using an optical map. The results, however, are not yet satisfying. RESULTS: To increase the accuracy of contig connections, we develop OMACC, which carefully takes into account length information in optical maps. Specifically, it rescales optical map and applies length constraint for selecting the correct contig sequences for gap closure. In addition, it uses an advanced graph search algorithm to facilitate estimating the number of repeat copies within gaps between contigs. On both simulated and real datasets, OMACC achieves a <10% false gap-closing rate, three times lower than the ~27% false rate by FINISH, while maintaining a similar sensitivity. CONCLUSION: As optical mapping is becoming popular and repeats are the bottleneck of assembly, OMACC should benefit various downstream biological studies via accurately connecting contigs into a more complete genome. AVAILABILITY: http://140.116.235.124/~tliu/omacc