In our lab, we study a host of questions regarding the fundamental nature of mutation and the role of sexual recombination in evolution, in both the theoretical and empirical realms. Below are some of the research areas we explore.   

The mixability theory for the role of sexual recombination in evolution 

For nearly a century, theoretical research on the role of sexual recombination in biology evolution has been guided by a tacit assumption that sex should somehow facilitate the increase in the population mean fitness measure as defined in population genetic models, even though this measure does not explicitly represent biological structure. The mixability theory for the role of sexual recombination in evolution argues instead that, by shuffling the genes, sexual recombination shifts the "focus" of natural selection from favoring particular genetic combinations of high fitness to favoring genes that perform well across many different genetic combinations. Thus, the interaction between sex and natural selection generates modular genetic elements that are simple and robust and are able to form novel genetic interactions.


Research in our lab is currently funded by the Israel Science Foundation and by the John Templeton Foundation

Exemplifying the fact that research in evolution sometimes inspires developments in Computer Science, mixability theory served as a source of motivation for the development of a method for the training of deep learning networks, called "dropout," which generates robust units that perform well across different combinations of units. Dropout has been one of the key breakthroughs that allowed for the success of deep learning and thus the AI revolution of our times.  

Further reading:
-Livnat et al. A mixability theory for the role of sex in evolution. Proceedings of the National Academy of Sciences, USA, 105:19803-808, 2008.

-Livnat et al. Sex, mixability, and modularity. Proceedings of the National Academy of Sciences, USA, 107:1452-7, 2010. 
-Vasylenko et al. Sex: The power of randomization. Theoretical Population Biology, 129:41-53, 2019.

-Srivastava et al. Dropout: a simple way to prevent neural networks from overfitting. The Journal of Machine Learning Research, 15:1929-58, 2014.

Interaction-based Evolution

Since Darwin, there have been two main ways of thinking about the fundamental nature of mutation and thus about how evolution happens. One has been random mutation and natural selection. According to it, the causes of mutation include factors like radiation, replication errors and oxidative stress, and are not important for our understanding of how evolution happens at the fundamental level. The second view, Lamarckism, argues that the organism can respond with beneficial genetic change directly to its immediate environment, but has difficulties explaining evolution beyond limited cases.


Livnat's theory of Interaction-based Evolution argues instead that complex genetic influences affect the probabilities of origination of mutations in a mutation-specific manner, and that these complex genetic influences are shaped by gradual adaptive evolution and vice-versa. This theory is distinct from the modifier theory of mutation rates in that it argues for genetic influences on mutation rates that are complex and mutation specific. We are now testing this theory empirically using novel, advanced methodologies. Links to our results will be posted upon publication. 


The study of mutational mechanisms is important not only for evolutionary biology but also for other fields, such as cancer. Cancer is a disease caused by mutations, and a better understanding of the fundamental nature of mutation may lead to further insights on cancer. 

Further reading:

-Livnat, A. Interaction-based evolution, Biology Direct 8:24, 2013.
-Livnat, A. Simplification, innateness, and the absorption of meaning from context: How novelty arises from gradual network evolution. Evolutionary Biology, 44:145-189, 2017. 

Mixability theory is based on mathematical models that examine sexual recombination and natural selection together. Mutation is further explored by the theory of Interaction-based Evolution, described below.  

© 2018 By the Livnat Lab. Proudly created with