MicroTrait: Tools for parameterization of microbially-explicit trait-based models with microbiome data

Microbiome functional components have become increasingly abundant thanks to the routine application of sequence-based assays in microbial ecology. There is now a growing interest to understand how the microbiomes' components, typically of high complexity, interact in space and time to define the function of microbial systems and their feedbacks to their surrounding environments. This necessitates going beyond the assumed relationships between constituent taxa and the properties of microbial communities. Trait-based representations of microbial function represent a promising approach to compress microbial diversity while retaining sufficient information relevant to an ecological context. Given issues with cultivation efficiency, opportunities for trait parameterization for complex microbiomes are limited. However, increasing connections are being made between genome-level properties or organisms and phenotypic traits relevant to ecological fitness. These genome-derived 'traits' can be used to initialize and parameterize mechanistic trait-based models spanning a scale of complexities and used to explore how patterns of trait distribution and co-occurrence relate to phylogenetic and ecological context. We describe, MicroTrait, a curated database of protein families, associated profile hidden markov models, and workflows to derive metabolic, resource acquisition-utilization, stress tolerance, and life history traits that enables distillation of genome sequences into microbial trait matrices. The suite of traits included in MicroTraits have been chosen to specifically inform trait based ecosystem models, which we showcase their use in defining guilds of soil microbes and parameterization of microbially-explicit trait-based models.