(674g) Development and Application of Integrated Pipeline for the Modeling and Analysis of Microbial Communities in the Doe Systems Biology Knowledgebase

Authors: 
Henry, C. S., Argonne National Laboratory
Weisenhorn, P., Argonne National Laboratory
Faria, J. P., Argonne National Laboratory
Edirisinghe, J. N., Argonne National Laboratory
Taylor, R. C., Pacific Northwest National Laboratory
Song, H. S., Pacific Northwest National Laboratory
Bernstein, H. C., Pacific Northwest National Laboratory
Zucker, J., Pacific Northwest National Laboratory
Lindemann, S. R., Pacific Northwest National Laboratory
Arkin, A. P., University of California, Berkeley
Understanding the molecular interactions and ecological principles that guide the behavior of microbial communities is critical to numerous applications in medicine, in the environment, and in industry. Toward this end, we have developed a pipeline within the DOE Systems Biology Knowledgebase (KBase) that supports the development and analysis of community metabolic models starting from metagenomic data. This pipeline integrates 11 analysis steps including: metagenome assembly; binning assembled contigs by species; assessment of assembly and genome quality and completeness; genome annotation; species metabolic model reconstruction; merging of models into a community model; mapping RNA-seq reads to individual species; model gapfilling; flux balance analysis; comparison of flux and expression profiles; and simulation of growth phenotypes. All eleven steps are available within a user-friendly point-and-click interface at http://narrative.kbase.us.

We applied this pipeline to understand the ecology and trophic interactions occurring within three microbiome-based datasets: (i) a lab-constructed community comprised of the cyanobacterium Thermosynechococcus elongatus supporting the heterotrophic bacterium, Meiothermus ruber; (ii) ten naturally occurring highly coupled communities comprised of 2-3 species each; and (iii) a larger 18-species natural community comprising an epsomitic phototrophic microbial mat in Hot Lake, Washington. We demonstrate how the tools in KBase can integrate community Biolog profiles and transcriptomic profiles in our community models of these systems. Further, we show how our platform can predict the trophic interactions that occur among species in each of these systems. Overall, we find significant trophic interactions in all communities, involving numerous essential metabolites, including amino-acids, vitamins, and cofactors. We also find numerous cases of metabolic hand-off, particularly in our lab constructed and Hot Lake communities, which both include an autotrophic organism that performs carbon fixation functions for all other species in the community. We see how species form ecological networks of interactions where each species derives chemical energy from the nutrients in the environment, while stability is preserved through essential trophic interactions. We also see how changes in the nutrient content of the environment can disrupt these webs of trophic interactions.