Probabilistic models in biology

Predictive understanding of signaling network kinetics is essential for design of rational intervention strategies in diseases. However, inference of network parameters, such as reaction rates and species abundances, that govern intracellular signaling network dynamics is challenging due to parameter non-identifiability and cell-to-cell variability. We present a maximum entropy (ME) inference of the joint probability distribution over network parameters that reproduces snapshots of cell to cell variability in species abundances measured at multiple time points. We studied a network implicated in many diseases; phosphorylation of protein kinase B (Akt, pAkt) via the EGF/EGFR pathway. Our framework combined with experimental data on cell-to-cell variability in pAkt abundances allowed us to computationally reconstruct the experimentally inaccessible ensemble of trajectories of intracellular pAkt dynamics. Using the ensemble of trajectories, we could predict the fraction of cells that entered the cell cycle after prolonged exposure to EGF. The analysis allowed us explore sub-populations of cells that failed to commit to cell cycle.

I will also talk about our recent work on understanding the role of horizontal gene transfer (HGT) in the evolution of E. coli. A significant fraction of the core genomes of Escherichia coli strains comprise mobile elements integrated at various sites. We find that the entire core genome is continually exchanged by homologous recombination with genome fragments acquired from other genomes in the population. Evolutionary groups appear to exchange DNA preferentially within the same group but also with other groups to different extents. Entering DNA is often fragmented by restriction systems of the recipient cell, with surviving pieces replacing homologous parts of the recipient chromosome. Coevolving populations of phages that package genome fragments and deliver them to cells that have appropriate receptors are likely mediators of most DNA transfers, distributing variability throughout the species.