SPAdes: is there anything new we could develop?

Breakthroughs of the third generation sequencing technologies bring to life the ideas of their usage in metagenomic assemblies. Unfortunately, low and uneven coverage, closely related strains and hard repetitive structure... [ view full abstract ]

Breakthroughs of the third generation sequencing technologies bring to life the ideas of their usage in metagenomic assemblies. Unfortunately, low and uneven coverage, closely related strains and hard repetitive structure prevent existing tools from generating reasonable assemblies from long read data. Instead of assembling long reads themselves we suggest to use them for repeat resolution and strain separation on top of assembly graph constructed from NGS data. We present HybridMetaSPAdes – an adaptation of metaSPAdes that utilizes third generation sequencing data for metagenomic hybrid assembly and strain level deconvolution.

The existing bacterial gene prediction tools assume that putative genes are encoded within a single contig of a genome / metagenome assembly and therefore they have limitations with respect to predicting longer genes with repetitive domains that span multiple contigs. antiSPAdes is a module of SPAdes assembler designed for better recovery of long genes with domain structure. Such genes often have high biomedical importance since they can be a source of active natural products including antibiotics. antiSPAdes uses domain structure in graph simplification and repeat resolving procedures resulting in high-quality candidates for complete gene sequence. Currently antiSPAdes is capable to reconstruct genes of non-ribosomal peptide synthetases and polyketide synthetases.