Microbes such as bacteria and yeast are able to maintain a narrow distribution of cell sizes by regulating the timing of cell divisions. In rich nutrient conditions, bacteria such as E. coli divide faster than their chromosomes replicate (!), implying that cells maintain multiple ongoing rounds of chromosome replication. How these processes are coupled and controlled is a fundamental question in cellular biology. We have shown that ideas from statistical physics are helpful in deciphering this problem, and in particular lead to insights into the structure of the correlations between cell cycle parameters. The models also account for the tight coupling between size control and DNA replication. Results on budding yeast and archaea show similar behavior, suggesting that the principles involved may be prevalent in nature across different domains of life - for reasons yet to be elucidated.
How single-cell variability affects the population growth
How does cell-to-cell variability affect the population growth rate? The answer to this question will determine whether evolution will tend to suppress or enhance fluctuations. We revisited this long-standing problem, considering realistic models where cell size is controlled. We found that this profoundly affects the results, and in contrast to the dogma in the field, found that single-cell variability is often detrimental rather than beneficial to the population growth. In other scenarios, population diversity results from the asymmetric segregation of cellular resources. We discovered a phase transition between a regime where homogeneity is optimal to one where asymmetric division is favorable. Drawing on ideas from statistical physics allowed us to develop novel techniques for inferring fundamental information regarding growth from single-cell data, in particular extracting the population growth rate from single lineages.
Modeling gene expression
We are developing new models for gene expression, which account for the coupling between mRNA and protein production and cell growth. Within these models the numbers of RNA polymerases and ribosomes are explicitly considered as dynamic variables, leading to important consequences regarding the growth dynamics. The models predict a regime where cells grow exponentially with mRNA and protein levels scaling with cell volume, and another with growth linear in time. We have also shown that by analyzing the time trajectory of the protein levels one can infer the relative fractions of the intrinsic and extrinsic noise in gene expression.
We developed models accounting for the action of transcription factors leading to interactions between the various genes and have shown that the gene regulatory network will become unstable at a critical size. Analyzing the structure of biological networks, we found global motifs that stabilize the network dramatically.
Shape regulation in bacteria
How do microbes such as bacteria maintain their shapes? We are applying ideas from statistical mechanics and materials science to this interdisciplinary problem, in collaboration with our experimental colleagues. Previously we have shown, experimentally and theoretically, that mechanical stresses can strongly affect cell wall growth in bacteria. We are currently working on elucidating how such mechanical cues may aid bacteria in restoring their native forms when their shape is perturbed, and how the binding of proteins to membranes may act as curvature sensors. We are also developing models for processes such as bacterial lysis where the interplay of turgor pressure, cell wall and membrane physics lead to intriguing dynamics at the single-cell levels.
Image credit: Arnaud Chastanet.
The physics of glasses and random matrix theory
The interplay of disorder and interactions can lead to remarkable effects, such as a glassy phase - many systems in nature exhibit slow dynamics, aging and memory effects, on time scales ranging from seconds to days. Previously, we studied electron glasses, which are systems in which electrons exhibit these phenomena. More recently, we extended our approach to other, "non-conventional" glasses, such as crumpled thin sheets and other mechanical systems as well as models of microbial evolution.
We have also studied various classes of non-Hermitian random matrices associated with biological networks, motivated by neural networks and gene regulatory networks.
We anticipate openings at the intersection of physics, applied mathematics, and biology. The candidate should have a strong quantitative background. The position can start at any time. Qualified persons are advised to send their CV and names of references.
Harvard is an equal opportunity employer and encourages applications from under-represented groups such as women and minorities.