Sergey Kryazhimskiy


My Research

I am a postdoc in the Desai Lab at the Department of Organismic and Evolutionary Biology at Harvard University. I am trying to uncover regularities in how microorganisms adapt to their environment and understand how these regularities emerge from the biology of the cell.

I use experiments in budding yeast, theory, and data from natural populations to address the following questions:

  • How predictable are evolutionary outcomes at genetic and phenotypic levels?
  • How many physiologically and genetically distinct ways to adapt are there in different environments?
  • How do the mutations alter the effects of other mutations, i.e., what is the structure of epistasis?

I am currently working on two main projects.

1. How does adaptation depend on the starting genotype?

Microbes rapidly adapt to their environment, allowing them to escape the immune system, resist drugs, and undermine performance of synthetically engineered circuits. Predicting evolutionary trajectories is therefore a major problem in biology, with wide-ranging practical implications. Several recent studies (e.g., Burch and Chao 2000, Blount et al 2008Bloom et al 2010, Woods et al 2011, Hayden et al 2011) have shown that epistatic interactions among mutations (i.e., when the effect of a mutation depends on the presence of other mutations in the genome) dramatically affect the course of adaptation implying that evolution may be essentially unpredictable. On the other hand, evolutionary outcomes may be statistically predictable if mutations leading to extreme and irregular changes in adaptability (also called evolvability) are rare, while mutations leading to small and regular changes in adaptability are common. We directly tested this hypothesis by measuring the variation in adaptability between related genotypes in laboratory yeast populations.

We evolved 640 populations of budding yeast (see Figure below), measured how their fitnesses changed over time, sequenced the full-genomes of 104 evolved clones to identify beneficial mutations, and made targeted reconstructions of mutations to characterize the type and strength of epistasis among them. We find that the average rate at which a strain adapts (its adaptability) depends not on the specific mutations initially present in its genome but only on its initial fitness, with fitter strains adapting predictably slower than less fit strains. Surprisingly, both low- and high-fitness strains sample beneficial mutations from a common pool, i.e., we found no evidence for sign epistasis. The fact that high-fitness strains are less adaptable than low-fitness strains, despite having access to the same pool of beneficial mutations, is explained by the “diminishing returns” epistasis (Chou et al 2011, Khan et al 2011). Critically, we find that this epistasis is “global” in the sense that the effects of beneficial mutations depend only on the fitness of the genetic background but not on the identities of the mutations that led to that fitness.

The paper is now in review and is available here. This is work with Dan Rice, Elizabeth Jerison, and Michael Desai.

2. Can we predict effects of mutations from cellular metabolism?

This is a figure from the classic paper by Daniel Dykhuizen, Tony Dean and Dan Hartl called "Metabolic flux and fitness" (link). In this paper, which I highly recommend, the authors laid the foundation for understanding microbial fitness from biochemical principles. Building on their ideas, I am developing a theory for describing the effects of mutations and their interactions (epistasis) in terms of our current understanding of metabolism.