Back

A quantitative measure of functional redundancy in microbial ecosystems

Royalty, T. M.; Steen, A. D.

2020-04-23 microbiology
10.1101/2020.04.22.054593 bioRxiv
Show abstract

The concept of functional redundancy has received considerable attention in both the macroecology and microbial ecology literature. As a result, multiple metrics of functional redundancy have been proposed. These vary in how they weight trait levels, species abundance, functional richness, and species richness. Here we present a new functional redundancy metric tailored for community-aggregated traits, which are traits that are quantified at the community level and can be quantitatively partitioned among species. We call this metric Contribution Evenness (CE) because it measures how evenly species contribute to a community-aggregated trait. As CE is an evenness measurement, it ranges from 0 and 1, where 0 corresponds to a single species contributing to a community-aggregated trait and 1 corresponds to all species contributing equally. Using in silico simulations of species extinctions, we demonstrate that CE reflects the stability of an ecosystem function to species extinction, a hypothesized ecological consequence of functional redundancy. As a positive control and to illustrate how CE can be used with sequence data, we analyzed the functional redundancy of eight nitrogen-transforming pathways using 2,631 metagenome-assembled genomes from 47 marine sites. CE for marine nitrogen cycle marker genes was consistent with our qualitative understanding of which nitrogen pathways are most functionally redundant in the ocean. We found that, on average, the NH4+ assimilation pathway was the most functionally redundant (0.44 {+/-} 0.08) while dissimilatory nitrate reduction was the least redundant (0.005 {+/-} 0.005). As demonstrated here, CE provides a promising framework for measuring trait stability in microbiomes.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.