Current interests:
I am interested in understanding biological phenomena through general principles and theoretical models. My recent work has focused on how organisms adapt to changing environments. Different organisms exhibit morphological and behavioral traits, called phenotypes, that are characteristic of their species. However, even for the same species, the phenotypes vary among individuals and according to environmental conditions; moreover, some phenotypes are transmitted non-genetically across generations. My goal is to understand the role of such phenotypic variation and inheritance in biological adaptation and evolution.
Some patterns of phenotypic variation are common among different species. For example, a population of organisms may diversify into a mixture of individuals with distinct phenotypes, each being favorable for one or another environmental condition. Such phenotypic variation could help the population adapt to fluctuating environments, because there is always a subpopulation that is prepared for any environmental change. This phenomenon of phenotypic diversification is found in many organisms including bacteria, plants, and animals, which otherwise differ substantially in terms of genetic sequences or molecular mechanisms. Such a pattern of phenotypic variation is an adaptation strategy that is general in its functional aspect, rather than the mechanistic details that are idiosyncratic to each species. My work tries to elucidate the origins and functioning of such adaptation strategies.
- phenotypic diversification and evolutionary learning
Phenotypic diversification, also known as ``evolutionary bet-hedging'', has been widely studied in theory and experiments. It was established that bet-hedging provides the largest long-term growth benefit for a population if the distribution of phenotypes matches the statistics of environmental variation. To reach the optimal phenotype distribution, however, organisms must overcome an apparent mismatch in timescale: the individual lifespan is often too short for gathering necessary environmental statistics. In a theoretical work with Stanislas Leibler, I proposed an ``evolutionary learning'' process, by which the phenotype distribution of the population could be dynamically optimized (see this paper). It happens when each organism alters the phenotype distribution of its offspring to increase the frequency of its own phenotype, analogous to Hebbian learning in neuroscience. This process could be realized through various mechanisms of epigenetic inheritance, which are actively studied in biology.
- phenotypic plasticity and environment-to-phenotype mapping
Phenotypic plasticity is the ability of organisms to develop alternative phenotypes depending on environmental cues. It is an important aspect of evolution, because in such cases the environment plays not only a selective but also a formative role. Working with Pablo Sartori and Stanislas Leibler, I proposed to describe phenotypic plasticity by a mapping from environmental cues to phenotypic traits, which is an abstraction of many possible mechanisms for gene regulation. Using an optimized ``environment-to-phenotype mapping'', it was shown that the evolutionary benefit of phenotypic plasticity is determined by the predictability of future environment at the time of development (see this paper). Moreover, depending on the noise in environmental cues and the strength of natural selection, phenotypic plasticity can give rise to different forms of adaptation, including bet-hedging. The generality of this theoretical framework, including the possibility of incorporating epigenetic memory, is currently being studied.
Previous work:
- Protein structural patterns from sequence data
For an exploratory project with the advice of William Bialek and in collaboration with Jeremy England, Thierry Mora, and Aleksandra Walczak, I tried to extract statistical patterns that are common in the amino acid sequences of all proteins from a single organism, the bacterium E. coli. We inferred from the distribution of protein sequences a maximum entropy model with pairwise interactions between amino acids. Correlations were found at short distances, and did not vanish at long distances. These features were captured by a short-range model with translation-invariant interactions, and by a fully-connected model with distance-independent interactions. Our models were used to look for stable patterns of protein structure and diverse phases of protein composition.
- Selective Plane Illumination Microscope (SPIM)
For an experimental project done at the Shaevitz Lab @ Princeton, I built a prototype SPIM to construct 3D images of biological samples at micrometer resolution. Cylindrical lenses are used to create a thin light sheet near the focal point that illuminates a single layer across the sample, which is observed from the direction perpendicular to the light sheet. The light sheet then scans through the sample to take a series of images at different depths, which are used to construct 3D images. I used the microscope to image zebrafish embryo and could resolve single cells.