Abstract:
Studies of the human microbiome have brought paradigm-shifting implications for translational research and clinical care, and, is now recognized as significant across a range of human organ systems. Despite significant progress in the field over the last decade, a holistic analysis of bacteria, fungi and viruses (the "multi-biome") is rarely performed despite this most closely representing the true in-vivo state. Integration of these high- dimensional datasets brings challenges in terms of complexity and their translation into clinically actionable outputs. To address this "analytical bottleneck", we sought to build a computational pipeline for integration of bacterial, fungal and viral datasets from a single well characterised patient population (a process we coin "integrative microbiomics") as a proof of principle in work described below. Having successfully integrated bacterial, fungal and viral datasets, we characterise the integrated microbial components by identifying a statistically significant super-consensus network representing possible mathematical microbial interactions (which
we term the "interactome"). Further, we show that cross-talk between microbes is as significant as the isolated microbes not if higher, in driving specific disease states.
