Sequential co-assembly reduces computational resources and errors in metagenome-assembled genomes

Generating metagenome-assembled genomes (MAGs) from DNA shotgun sequencing datasets can demand considerable computational resources. Here we describe a sequential co-assembly method that reduces the assembly of duplicate reads through successive application of single-node computing tools for read assembly and mapping. Using a simulated mouse microbiome DNA shotgun sequencing dataset, we demonstrated that this approach shortens assembly time, uses less memory than traditional co-assembly, and produces significantly fewer assembly errors. Applying sequential co-assembly to shotgun sequencing reads from (i) a longitudinal study of gut microbiomes from undernourished Bangladeshi children and (ii) a 2.3 terabyte dataset, generated from gnotobiotic mice colonized with pooled microbiomes from these children, that was too large to be handled by a traditional co-assembly approach, also demonstrated significant reductions in assembly time and memory requirements. These results suggest that this approach should be useful in resource constrained settings, including in low- and middle-income countries.