Simulation-based approaches to characterize the effect of sequencing depth on the quantity and quality of metagenome-assembled genomes

Abstract

We applied simulation-based approaches to characterize how microbial community structure influences the amount of sequencing effort to reconstruct metagenomes that are assembled from short read sequences. An initial analysis evaluated the quantity, completion, and contamination of complete-metagenome-assembled genome (complete-MAG) equivalents, a bioinformatic-pipeline normalized metric for MAG quantity, as a function of sequencing effort, on four preexisting sequence read datasets taken from a maize soil, an estuarine sediment, the surface ocean, and the human gut. These datasets were subsampled to varying degrees of completeness in order to simulate the effect of sequencing effort on MAG retrieval. Modeling suggested that sequencing efforts beyond what is typical in published experiments (1 to 10 Gbp) would generate diminishing returns in terms of MAG binning. A second analysis explored the theoretical relationship between sequencing effort and the proportion of available metagenomic DNA sequenced during a sequencing experiment as a function of community richness, evenness, and genome size. Simulations from this analysis demonstrated that while community richness and evenness influenced the amount of sequencing required to sequence a community metagenome to exhaustion, the effort necessary to sequence an individual genome to a target fraction of exhaustion was only dependent on the relative abundance of the corresponding organism and its genome size. A software tool, GRASE, was created to assist investigators further explore this relationship. Re-evaluation of the relationship between sequencing effort and binning success in the context of the relative abundance of genomes, as opposed to base pairs, provides a framework to design sequencing experiments based on the relative abundance of microbes in an environment rather than arbitrary levels of sequencing effort.