Final Report Summary - IR-DC (Individual Robustness in Development and Cancer)
Biological systems are robust to perturbations, with many genetic, stochastic and environmental challenges having no or little phenotypic consequence. However, the extent of this robustness varies across individuals, for example the same mutation or treatment may only affect a subset of individuals. The overall objective of this project was to better understand cellular and molecular mechanisms that confer this robustness and its variation across individuals.
To better understand why mutations have variable outcomes we used deep mutational scanning to comprehensively analyze how mutations combine to affect diverse biological processes including individual proteins and RNAs, regulatory interactions and networks. We also developed open source software to analyze these datasets. We found that using both statistical and mechanistic models we could accurately predict how multiple mutations combine to alter molecular phenotypes. We also developed a method that uses mutagenesis alone to determine the 3D structures of proteins, including as they are performing important functions inside a cell.
To understand robustness to mechanical perturbations we developed single cell tracking methods to follow the divisions and movements of all cells in the early C. elegans embryo. We found that specific regulative cell movements compensate for the effects of compression in this embryo, conferring robustness to a major physiological perturbation.
Somatic (non-inherited) mutations are a striking example of stochastic processes in biology and the most important cause of cancer. We discovered that differential DNA repair is the major cause of somatic mutation rate variation across the human genome and also that carcinogens like alcohol elevate the mutation rate specifically in the most important active regions of the genome because of the engagement of ‘error-prone’ processes that repair DNA. We used the thousands of mutations in cancer genomes to learn the rules for when mutations introducing a ‘stop’ signal into an mRNA result in its destruction and we applied these rules to better understand cancer genes, to improve gene editing and to better understand why some patients respond better to cancer immunotherapy. Finally, we developed statistical methods that allowed us to determine how different mutations interact in cancer, which allowed us to identify new rare cancer predisposition genes.
To better understand why mutations have variable outcomes we used deep mutational scanning to comprehensively analyze how mutations combine to affect diverse biological processes including individual proteins and RNAs, regulatory interactions and networks. We also developed open source software to analyze these datasets. We found that using both statistical and mechanistic models we could accurately predict how multiple mutations combine to alter molecular phenotypes. We also developed a method that uses mutagenesis alone to determine the 3D structures of proteins, including as they are performing important functions inside a cell.
To understand robustness to mechanical perturbations we developed single cell tracking methods to follow the divisions and movements of all cells in the early C. elegans embryo. We found that specific regulative cell movements compensate for the effects of compression in this embryo, conferring robustness to a major physiological perturbation.
Somatic (non-inherited) mutations are a striking example of stochastic processes in biology and the most important cause of cancer. We discovered that differential DNA repair is the major cause of somatic mutation rate variation across the human genome and also that carcinogens like alcohol elevate the mutation rate specifically in the most important active regions of the genome because of the engagement of ‘error-prone’ processes that repair DNA. We used the thousands of mutations in cancer genomes to learn the rules for when mutations introducing a ‘stop’ signal into an mRNA result in its destruction and we applied these rules to better understand cancer genes, to improve gene editing and to better understand why some patients respond better to cancer immunotherapy. Finally, we developed statistical methods that allowed us to determine how different mutations interact in cancer, which allowed us to identify new rare cancer predisposition genes.