Inferring how circadian rhythms drive cellular proliferation in development and disease: Single cells in our bodies can track external time based on 24-hour oscillations of “circadian” proteins. Aberrations in these oscillations can result in susceptibility to cancer and tumor growth, suggesting an intricate connection between the circadian clock and cellular proliferation. In this research program, we are investigating how the circadian clock controls proliferation and cell fate decisions in cancer. Using long term time-lapse microscopy and Fluorescence In Situ Hybridization, we quantify the levels of circadian and cell-cycle proteins/mRNA in single cancer cells and how they change in the presence of anti-cancer therapies. We aim to infer the nature of coupling between the circadian oscillations and the cell-cycle from this data by developing computational approaches rooted in Bayesian inference and probabilistic graphical models. 

A related paper can be found here:  https://www.nature.com/articles/s41467-018-07788-5

 

 

Elucidating principles and timing of cell-fate decisions in cancer:  Apparently identical cancer cells exhibit a puzzling ‘fractional killing’ response to anti-cancer agents, where only a subset of cells die. A quantitative understanding of this phenomenon for specific agents such as chemotherapy or radiation, is currently missing. The goal in our lab is to predict population level behavior of cancer cells based on theoretical models of single cell dynamics and cell fate decisions, using time lapse imaging data of cancer cells with various protein reporters. These projects will combine ideas from mathematical biology (theory of age-structured populations), with machine learning techniques (factor graphs and message passing algorithms) and time lapse fluorescence microscopy of cancer cells.

A related paper can be found here:  https://www.nature.com/articles/s41467-018-07788-5

 

 

Designing improved dosing strategies for combination therapy in cancer: Developments in our ability to deeply sequence the genome and interrogate non-genetic cellular heterogeneity are rapidly improving our understanding of the origins of drug resistance and patient-to-patient variability in cancer. However, driven largely by theoretical and computational modeling, it has also become evident that current dosing regimens of anti-cancer therapies are far less effective at dealing with drug resistance than they could potentially be. Additional variables like the body’s circadian clock are increasingly being recognized as crucial factors in driving heterogeneous patient responses to treatments. To incorporate all these variables into a predictive framework, we combine evolutionary models of cancer progression with in vitro experiments on cancer cells to design improved strategies for controlling cancer cell proliferation. These strategies can then be tested in mouse models and eventually in clinical trials. 

A related paper can be found here:  https://cancerres.aacrjournals.org/content/77/14/3908.short

 

 

Leveraging DNA/histone modification patterns to understand epigenetic inheritance and develop tools for early cancer detection: Cells of various tissue types such as brain, lung, liver etc comprise identical DNA, yet they are morphologically dissimilar and perform distinct tasks. This phenomenon is partly achieved via chemical modifications like methylation on DNA and histones, which ensure that the DNA is ‘read’ differently in different cell types. The stability of these ‘epigenetic’ modifications is dependent on the dynamics of their deposition and erasure from the DNA and histones, both during and in between cell division events. How and with what fidelity these modifications are maintained and transmitted across cellular generations remains an incompletely understood question. We will combine bioinformatics (analysis of DNA and histone methylation, HiC and replication timing datasets) with methods from statistical physics and inference techniques to quantitatively understand mechanisms of epigenetic inheritance and develop novel biomarkers for early cancer detection from cell-free DNA in patient blood samples.

A related paper can be found here:  https://www.sciencedirect.com/science/article/pii/S1097276520300381