Alladi Ramakrishnan Hall

Computational approaches to decoding microbial interactions microbiomes

Karthik Raman

Department of Biotechnology, IIT Madras

Metabolic interactions are known to drive the organisation of various microbiomes. It is of great interest to study the possible interactions between microbes in communities, understand the keystone species in these microbiomes, and how they influence one another and ultimately shape the structure of the microbiome. In this talk, I will focus on computational approaches we have developed, to understand the organisation of a variety of microbiomes, focussing on their metabolic networks. We have used complementary approaches from graph theory and constraint-based modelling to systematically study microbiomes ranging from those in the human gut and the eye, to extreme environments such as those aboard the International Space Station and deep-sea hydrothermal vents. In each of these environments, we identify unique interaction patterns and possible metabolic exchanges and dependencies amongst the organisms. We have developed a suite of novel computational approaches to study these microbial communities and understand key organisms and dependencies underlying these complex systems. Metabolic modelling, through a combination of graph-theoretic approaches and steady-state constraint-based modelling, paints a more comprehensive picture of possible microbial interactions, which are as yet inscrutable to this extent by experimental approaches. Our results point toward key dependencies of microorganisms in various environments. Our approach also underscores the importance of complementary modelling approaches in dissecting a fairly complex microbiome and understanding various possible interactions. Our methodology is fairly generic and can be readily extended to predict microbial interactions in other interesting milieu and generate testable hypotheses for wet lab experiments.