Publicação
Evolution of modularity in biological signalling networks
| Resumo: | The concept of modularity associated to signalling and gene expression regulatory networks has been a field of study for the past 20 years. There are a few theories that try to explain the origin of this architecture in biological networks but there is no agreement on which factors contribute more and are actually responsible for modularity to arise as a predominant topology. One of the ideas that is most accepted in the scientific community is that if a population of organisms is exposed between every few generations to new environmental challenges, as long as they have common sub-problems, that should be enough for modularity to be selected as the preferred network architecture type. In this work, we try to approach this problem in a different fashion, where our digital organisms population is represented by a set of directed graphs and they were exposed to different environments during their lifetime. Our objective is to prove that this condition alone is enough for modularity to arise on the population's signalling networks. Also we evaluate the influences of mutational parameters and their individual contributions, as well as the impact in terms of number of environments and how similar they are on the evolution towards modular networks. Our results show that it is possible to have an evolution towards modular networks just by exposing organisms to different environmental challenges through their life time, and not necessarily with environments with common sub-problems. Additionally, faster gene duplication rates and a slower gene interactions mutation rates are important in this process. Gene elimination does not seem to have any impact. This work also shows that fitness and modularity are not directly correlated, even if modularity may represent an evolutionary advantage, their evolution patterns can be different. Notably, the same simulation conditions that makes possible for modularity to arise, can also produce high fit populations that are not modular. |
|---|---|
| Autores principais: | Jorge, Daniel Vilar |
| Assunto: | Teses de mestrado - 2015 |
| Ano: | 2015 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório da Universidade de Lisboa |
| Resumo: | The concept of modularity associated to signalling and gene expression regulatory networks has been a field of study for the past 20 years. There are a few theories that try to explain the origin of this architecture in biological networks but there is no agreement on which factors contribute more and are actually responsible for modularity to arise as a predominant topology. One of the ideas that is most accepted in the scientific community is that if a population of organisms is exposed between every few generations to new environmental challenges, as long as they have common sub-problems, that should be enough for modularity to be selected as the preferred network architecture type. In this work, we try to approach this problem in a different fashion, where our digital organisms population is represented by a set of directed graphs and they were exposed to different environments during their lifetime. Our objective is to prove that this condition alone is enough for modularity to arise on the population's signalling networks. Also we evaluate the influences of mutational parameters and their individual contributions, as well as the impact in terms of number of environments and how similar they are on the evolution towards modular networks. Our results show that it is possible to have an evolution towards modular networks just by exposing organisms to different environmental challenges through their life time, and not necessarily with environments with common sub-problems. Additionally, faster gene duplication rates and a slower gene interactions mutation rates are important in this process. Gene elimination does not seem to have any impact. This work also shows that fitness and modularity are not directly correlated, even if modularity may represent an evolutionary advantage, their evolution patterns can be different. Notably, the same simulation conditions that makes possible for modularity to arise, can also produce high fit populations that are not modular. |
|---|