With the lack of clear identities for peaks, and with many identified peaks lacking known pathways, analysis is often limited to correlations alone. In this talk, we will discuss recent advances in tools and workflows in KBase and ModelSEED that are expanding the possibilities and opportunities for the use of metabolic models to integrate multi-omics data for the discovery of novel biochemical pathways.

Specifically, we have made significant improvements to our pipeline for the rapid reconstruction of metabolic models from sequence data, including isolate genomes and metagenomes. Now models have hundreds of additional genes and reactions, produce energy in biologically relevant ways, and include tailored templates for archaea, bacteria, plants, fungi, and cyanobacteria. We also offer a fully integrated pipeline for the prediction of novel biochemical compounds and reactions using cheminformatics approaches, including prediction of novel promiscuous enzymatic reactions and spontaneous chemical reactions. Finally, we have flux balance analysis workflows for combining genomic-based and novel chemical networks together to predict pathways to explain metabolomics data.

Scientifically, we will explore how these improved tools permit us to study pathway variation across the microbial tree of life, learn insights about microbial diversity and variation from microbiome data, and study evolutionary implications over how potential spontaneous reactions occur across the known metabolic pathways. We’ll demonstrate our multi-omics integration tools to discover new pathways in the JCVI minimal genome and to mechanistically map metabolites to microbes within the human microbiome. Our exploration of microbiome data demonstrates organizing principles for the assembly and function of microbiome systems.