Metabolism is a complex biological system involving many metabolites, reactions, and enzymes that work together for a defined biological function. Parasites and hosts compete for the same metabolic resources in the host environment to ensure survival. They attempt to optimally allocate metabolic resources for sustained proliferation and combat. In this project, we build / tailor systems-level microeukaroytic parasite and mammalian host GEMs by performing annotation of an organism’s genome, a level of detailing many biological pathway databases lack. Using constraint-based flux analyses of these GEMs complemented with 'omics'-based tailoring we perform optimal pathway tracing, estimate optimal metabolite synthesis, infer essential metabolic reactions and test for regulatory effects.

Techniques: Customized genome/protein annotation pipelines, genome-scale metabolic reconstructions, constraint-based flux analyses, protein large-language models

Current parasites of interest: Aspergillus fumigatus, Cryptococcus neoformans, Cryptosporidium parvum

During infection, the interaction of innate immune cell receptors with pathogen-activated molecular patterns triggers differentially-timed signal transduction responses for cytokine/chemokine production, recruitment of adaptive immune cells to the site of infection, production of antimicrobial peptides & complement factors. Understanding the link between signal transduction, transcription, and host immunometabolism is crucial for developing immunomodulatory therapies or vaccines. Similarly, parasite lifecycle, oxygen availability, nutrient availability, drug susceptibility and host factors modify gene regulatory patterns to generate infection phenotypes. The aim of this project is to develop/apply machine – learning based models for the prediction of transcription factor – target interactions using transcriptomics datasets thereby obtaining genome-scale transcriptional regulatory networks.

Techniques: Statistical / machine learning based inference, TF-binding site prediction, co-expression networks, differential regulatory network analyses

Apart from systems-level analyses, the development of automated systems biology model tools and pipelines is pivotal for advancing metabolic network reconstruction, visualization, metabolic engineering, functional annotation of enzymes / reactions (molecular, subcellular and process-level functions) and kinetic parameterization. We aim to also curate harmonized databases of reaction rules, metabolite structures (SMILES), LLM-derived features to enable automated gap-filling and improvement of metabolic reaction annotation. Similarly, we are also interested in the development of pathway visualization, metabolic engineering and analyses toolboxes for interactively exploring pathway topologies, flux distributions, intervention targets and facilitating de novo pathway design. The pipelines will be deployed as user-friendly web apps with APIs for DBTL workflows, standalone tools with active updates.

Genomics features like single nucleotide polymorphisms (SNPs), copy number variations (CNVs), horizontal gene transfers (HGTs) in bacterial strains are linked to antimicrobial resistance (AMR) mechanisms by evolving drug metabolism functions, drug efflux mechanisms and alternative pathway functions to ensure survival. Integrating such features into genome-scale metabolic reconstructions can help us assess impacts on network topology (e.g., connectivity, modularity), altered metabolic phenotypes and enrichment of machine learning models for AMR prediction. Expanding this concept to human genome cohorts which catalog population-specific SNPs, we envision data-driven modeling drug susceptibility, host-pathogen metabolic interactions and predisposition of individuals to diseases.