Mathematical Biology and Bioinformatics Laboratory

Gene networks control organism phenotype in health and disease by dynamically processing information introduced by genotype and environment. We are rapidly entering an era of “personal genomics” wherein data on gene structure and expression will be available for large numbers of individuals. Such experiments demonstrated the existence of abundant variation in human and other organism individual DNA, however a fundamental challenge, critical for realizing the promise of personalized medicine, is to understand how and why some individual variations lead to phenotypic changes, while others do not. It is straightforward to use the systems biology approach to answer on these questions, as systems biology studies both structure and dynamical behavior of genetic networks. We interrogate the interplay between variability and robustness at different scales, namely in the context of different network motifs as well as small and well-studied genetic networks controlling the body plan formation in fruit fly. Our goal is to understand in which situations genetic variability leads to alteration of molecular phenotype and in which it is buffered due to mutual interactions of network genes.

We model transcriptional gene regulation by using a hybrid approach and applying it to the well characterized segmentation gene regulatory networks in Drosophila development. This approach incorporates methods of statistical thermodynamics to calculate the activation probability of the target gene taking into account information about its regulatory genetic sequence (enhancers) and the transcription factors regulating the gene expression. It is important to understand at this level how a specific molecular configuration of the regulatory sequence gets formed and, being formed, how it influences the basal transcriptional machinery initiating transcription. Modelling stochastically the main processes leading a transcription factor molecule to its specific site on the DNA, we aim to investigate formation of the enhancer molecular configurations in terms of the realistic statistical distribution of probabilities of all possible configurations. This distribution is then linked to the probability of the target gene activation by assigning specific statistical weights to different configurations of the enhancer, and derivation of these weights is another essential direction of the research. At the next level of the hybrid modelling approach, we use dynamical equations of the reaction-diffusion type to describe the spatiotemporal patterns of the target gene expression based on the calculated probability of its activation. The ultimate goal of this approach is to formulate a quantitative map between the genotype (regulatory genetic sequence) and the corresponding molecular phenotype (gene expression level) with the single-nucleotide precision. As a model genetic system, we use the Drosophila gap genes, which is a class of segmentation genes controlling the body plan formation during the early embryo development.

We study mechanisms of intracellular vesicular transport and investigate the influence of the spatial-temporal localization of components of signaling cascades onto their functioning. We emphasize close coupling between experiments, computations, and theory, and use the cell cultures as an experimental system for model validation. Our work includes development of algorithms and software for quantitative analysis of the vesicular transport in cells; acquisition of quantitative data about vesicle trajectories, design of electronic atlas of standardized temporal images of vesicle trajectories, development of a mathematical model of vesicle transport. We also study the influence of signaling proteins in different cell compartments onto the functioning of the EGFR signaling pathway as a whole.