Purushottam Dixit, PhD
Assistant Professor
Contact Information
My scientific background is in chemical engineering, biophysics, and systems biology. My research expertise is in the development of principled probabilistic computational tools for modeling dynamics of cellular networks. We are also developing models of evolutionary and ecological dynamics of bacterial populations. Another focus in my laboratory is in developing mechanistic dimensionality reduction methods to analyze biological data. During the last decade, I have studied dynamics, stochasticity, and heterogeneity of a wide range of biological systems from structural dynamics of biomolecules, to dynamics and heterogeneity in signaling and metabolic networks (Dixit et al. Cell Systems 2020, Lyashenko*, Niepel*, Dixit* et al. eLife 2020, Li*, Ji*, Dixit*, et al. Nature Metabolism, 2022, Goetz*, Akl*, and Dixit, eLife, 2024, Goetz et al. eLife, 2025), to evolutionary and ecological dynamics of bacteria (Dixit et al. PNAS, 2015, Dixit et al. Genetics, 2017, Ji*, Sheth*, Dixit*, et al, Nature Methods, 2019, Shahin, Ji, and Dixit, npj Systems Biology and Applications, 2023). We have also pioneered the development of inference tools using the principle of maximum path entropy to infer dynamics of biological systems from limited data (reviewed by us in Dixit et al. J. Chem. Phys. 2018 and Ghosh*, Dixit*, et al. Ann Rev. Phys. Chem, 2020).
What do you do with Data Science?
I have a deep interest in inverse problems in biophysics and systems biology. Here, while the mapping from data (e.g. abundances of mRNAs or proteins) to models (e.g. parametrized chemical reaction networks) is not unique, the underlying physics (dynamical equations, constraints, measurements) can impose limits on possible models. We have been developing approaches to solve inverse problems using the maximum entropy principle. In computational biophysics, we have pioneered the development of inference of Markov chains from limited data using the principle of maximum path entropy (reviewed by us in Dixit et al. J. Chem. Phys. 2018 and Ghosh*, Dixit*, et al. Ann Rev. Phys. Chem, 2020). We have also developed physics-inspired approaches to approximate arbitrary probability distributions (Dixit, Phys. Rev. Res., 2020).