About

Michael Desai is a Professor of Organismic and Evolutionary Biology and Physics at Harvard University. His research focuses on using both theory and experiments to study evolutionary dynamics and population genetics, particularly in situations where natural selection is pervasive. His work aims to deepen understanding of how evolutionary processes operate under various conditions, contributing to the fields of molecular and cellular biology, as well as physics.

Research topics

• Biology
• Genetics
• Evolutionary biology
• Neuroscience
• Virology
• Computational biology
• Biochemical engineering
• Internal medicine
• Engineering
• Medicine
• Cell biology

Selected publications

• Unifying theories in high-dimensional biophysics: approaches, challenges and opportunities

  npj Systems Biology and Applications · 2026-03-28

  articleOpen access

  Across biological subdisciplines, the last decade has seen an explosion of high-dimensional datasets. At the ICTS workshop ‘Unifying Theories in High-Dimensional Biophysics’, we discussed whether this high dimensionality poses a challenge or an opportunity for theoretically describing, understanding and predicting biological systems. We discussed methods, models and frameworks that can be used for this purpose. This Comment summarizes our discussions from the perspectives of individual participants.

  PublisherOA PDFDOI

• Sex decreases the pleiotropic costs of local adaptation by purging hitchhiking load

  Science · 2026-04-23

  articleSenior authorCorresponding

  Understanding the evolutionary mechanisms that maintain sex despite its direct costs is a long-standing challenge. Previous work has shown that sexual recombination can accelerate adaptation, in part by separating beneficial mutations from deleterious hitchhikers. However, earlier studies focused on effects of sex in a constant environment. We show that recombination provides an advantage during changing conditions in promoting the evolution of generalist phenotypes by reducing pleiotropic costs caused by local adaptation. Using laboratory evolution in Saccharomyces cerevisiae , we show that hitchhiking load leads to pleiotropic costs and hence specialization in response to local adaptation in asexual but not sexual lineages. This provides evidence that sex can be maintained over long evolutionary timescales because it enables lineages to persist in the face of environmental change.

  PublisherDOI

• Dynamics of Contaminant Microbes in Bioethanol Production from Sugarcane

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-02-04

  articleOpen access

  ABSTRACT The dynamics and impact of microbial contaminants in industrial sugarcane bioethanol production in Brazil were investigated through a two-year metagenomic study across two biorefineries. Shotgun metagenomic sequencing revealed that temporal shifts in the contaminant microbiome dynamics within production seasons were more pronounced than inter-annual or inter-mill variations. While Saccharomyces spp. dominated, bacterial communities, primarily within the Firmicutes phylum and dominated by the genera Lactobacillus , Limosilactobacillus , and Bacillus , exhibited dynamic changes. Correlation analyses with industrial process parameters revealed a complex interplay: lower Lactobacillus levels in one mill were associated with increased ethanol yield, whereas higher levels in another mill correlated with reduced yeast viability and increased flocculation. The presence of Limosilactobacillus was linked to decreased yeast viability, whereas Bacillus showed potential for inhibiting both Lactobacillus and Limosilactobacillus . These findings highlight the nuanced and species-specific impacts of bacterial contaminants on bioethanol production, underscoring the need for strain-level functional studies and targeted interventions to optimize fermentation efficiency and stability in industrial settings.

  PublisherOA PDFDOI

• Laboratory yeast crosses reveal limited epistasis in the genetic basis of complex traits

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-04-06

  articleOpen accessSenior authorCorresponding

  Abstract Mapping the genetic basis of complex traits is complicated by the presence of epistatic interactions between loci. While work in molecular genetics identifies numerous specific genetic interactions, statistical analyses of quantitative traits frequently conclude that additive (nonepistatic) models explain most heritable variation. However, these conclusions are typically limited by the narrow range of genetic relatedness(e.g. in F1 offspring of a biparental or circular cross). Here, we use a barcoded panel of Saccharomyces cerevisiae genotypes with a broad range of relatedness to quantify the effects of epistasis on the genetic architecture of seven complex traits. We find limited contributions of epistasis to the genetic basis of these traits. These results indicate that epistasis beyond that detected in standard yeast crosses may exist, yet it contributes little to phenotypic variance in these systems.

  PublisherDOI

• Inferring genotype–phenotype maps using attention models

  PNAS Nexus · 2026-02-27

  articleOpen access

  Predicting phenotype from genotype is a central challenge in genetics. Traditional approaches in quantitative genetics typically analyze this problem using methods based on linear regression. These methods generally assume that the genetic architecture of complex traits can be parameterized in terms of an additive model, where the effects of loci are independent, plus (in some cases) pairwise epistatic interactions between loci. However, these models struggle to analyze more complex patterns of epistasis or subtle gene-environment interactions. Recent advances in machine learning, particularly attention-based models, offer a promising alternative. Initially developed for natural language processing, attention-based models excel at capturing context-dependent interactions and have shown exceptional performance in predicting protein structure and function. Here, we apply attention-based models to quantitative genetics. We analyze the performance of this attention-based approach in predicting phenotype from genotype using simulated data across a range of models with increasing epistatic complexity, and using experimental data from a recent quantitative trait locus mapping study in budding yeast. We find that our model demonstrates superior out-of-sample predictions in epistatic regimes compared to standard methods. We also explore a more general multienvironment attention-based model to jointly analyze genotype-phenotype maps across multiple environments and show that such architectures can be used for "transfer learning"-predicting phenotypes in novel environments with limited training data.

  PublisherOA PDFDOI

• Parallel but distinct adaptive routes in the budding and fission yeasts after 10,000 generations of experimental evolution

  Nature Ecology & Evolution · 2026-03-13

  articleOpen access

  Experimental evolution has been a useful tool for investigating long-term temporal evolutionary dynamics and molecular mechanisms underlying adaptation. However, extracting fundamental principles and predictive features of evolutionary outcomes from these datasets remains challenging. Here we sought to circumvent these challenges by comparing distant yeast species that share several evolutionary features but differ in evolutionary history and genome architecture, that is Saccharomyces cerevisiae and Schizosaccharomyces pombe. We evolved ten populations of the fission yeast for 10,000 generations in the same conditions as a pre-existing budding yeast dataset, allowing us to observe repeatable evolutionary outcomes within species but diverse molecular targets of adaptation across species. The most frequent route of adaptation was through changes in carbon flux metabolism, which was previously unseen in S. cerevisiae evolved populations, but similar evolutionary paths have been observed in wild populations. This suggests that parallelism is pervasive and that mechanisms of adaptation can be shared among closely related or distant species. Despite similar gene content and identical environments, recurrent adaptation across S. pombe populations involved different genes than in S. cerevisiae and was detectable mostly at the transcriptomic level. This indicates that trans-regulatory effects and contingency may contribute to differences in evolutionary outcomes between these species.

  PublisherOA PDFDOI

• Epistasis and background dependence in the evolution of Omicron variants of the SARS-CoV-2 Spike protein

  Molecular Biology and Evolution · 2025-11-19 · 3 citations

  preprintOpen accessSenior authorCorresponding

  The rapid and repeated emergence of SARS-CoV-2 variants, particularly within the Omicron lineage, highlights the virus's remarkable ability to adapt under shifting immune pressures. A central molecular battleground in this evolutionary arms race is the spike receptor-binding domain (RBD), which must simultaneously maintain high affinity for the human ACE2 receptor while evading recognition by neutralizing antibodies. In this study, we construct and analyze multiple combinatorial libraries of SARS-CoV-2 RBD variants spanning major branches of Omicron evolution, including BA.1, BA.2, BA.5, XBB, and JN.1. Using high-throughput yeast display and binding assays, we map the effects of thousands of mutations and their combinations on ACE2 binding and antibody evasion. Our results reveal that while many RBD mutations exhibit additive effects, several mutations interact epistatically in a background-dependent manner. In particular, we identify synergistic interactions between BA.1 and BA.5 mutations that enhance antibody evasion, likely facilitating the rise of recombinant variants and convergent evolution. Conversely, some mutations show lineage-restricted compatibility, suggesting potential constraints on future evolutionary trajectories. Our comprehensive genotype-to-phenotype maps uncover both rugged and smooth regions of the viral fitness landscape and underscore the importance of epistasis in shaping SARS-CoV-2 evolution. These findings improve our ability to anticipate future viral variants and provide a framework for understanding how host-pathogen co-evolution unfolds at the molecular level.

  PublisherDOI

• Experimental evolution in an era of molecular manipulation

  Nature Reviews Genetics · 2025-07-21 · 12 citations

  reviewSenior author
  PublisherDOI

• Impact of fluctuating environments on the fitness and robustness of evolving laboratory and industrial <i>Saccharomyces cerevisiae</i> strains

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-12-06

  articleOpen accessSenior author

  Abstract Adaptation occurs through the selection of beneficial mutations enhancing fitness in a specific environment. However, since environments vary across time and space, mutations that are positively selected in one environment may be less beneficial or detrimental in others. Here, we investigate the evolution of microbial robustness (i.e. a consistent fitness across many diverse environments) through the adaptive evolution of two genetically distinct Saccharomyces cerevisiae populations in fluctuating conditions, followed by fitness assays and whole-genome sequencing. Our results indicate that the haploid laboratory strain S288C achieved higher average fitness than its parental strain, particularly when evolved in fluctuating environments compared to constant environments, but did not show increased robustness. In contrast, populations of the industrial diploid strain Ethanol Red failed to achieve significant fitness improvement under both fluctuating and constant evolution regimes but became more robust. Populations that adapted to fluctuating conditions acquired mutations in genes involved with cell morphology and protein degradation. Overall, our results emphasise the importance of parental traits in shaping fitness and robustness during adaptive laboratory evolution.

  PublisherDOI

• Sex decreases the pleiotropic costs of local adaptation

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-03-31 · 1 citations

  preprintOpen accessSenior authorCorresponding

  Abstract Understanding the evolutionary mechanisms that maintain sex throughout nature despite its substantial direct costs is a longstanding challenge in biology. Previous work has shown that sexual recombination provides a key advantage in speeding adaptation, in part by separating beneficial mutations from deleterious hitchhikers. However, these earlier studies have focused on the effects of sex in a constant environment. Here, we show that recombination also provides a key advantage in fluctuating conditions, promoting the evolution of generalist phenotypes by reducing the pleiotropic costs of local adaptation. Using laboratory evolution in S. cerevisiae as a model system, we show that hitchhiking genetic load leads to pleiotropic costs and hence specialization in response to local adaptation in asexual but not in sexual lineages. This provides the first direct evidence that sex can be maintained over longer evolutionary timescales because it enables lineages to persist in the face of environmental change.

  PublisherOA PDFDOI

Recent grants

• The Evolution of Evolvability in Microbial Populations

  NSF · $798k · 2020–2025

• Recombination and the Dynamics of Adaptation in Experimental Saccharomyces Cerevisiae (yeast) Populations

  NSF · $669k · 2017–2020

• Microbial Adaptation and the Statistics of Epistasis and Pleiotropy

  NIH · $4.5M · 2013–2027

• Interference Selection and the Statistics of Evolution in Microbes

  NSF · $480k · 2013–2016

Frequent coauthors

• Benjamin H. Good

  Chan Zuckerberg Initiative (United States)

  68 shared

• Alex N. Nguyen Ba

  University of Toronto

  56 shared

• Aleksandra M. Walczak

  Centre National de la Recherche Scientifique

  48 shared

• Elizabeth R. Jerison

  45 shared

• Angela M. Phillips

  University of California, San Francisco

  40 shared

• Milo S. Johnson

  University of California, Berkeley

  37 shared

• Katherine R. Lawrence

  Oxford Nanopore Technologies (United Kingdom)

  37 shared

• Sergey Kryazhimskiy

  University of California, San Diego

  29 shared