This is a resume of the workshop on Evolutionary Systems Biology which took place a month ago in Edinburgh. At last I have the time to post it here in my blog 🙂
Have a look at my previous general post on this workshop, to see what it is about. I am sorry I could not take notes on all the talks I attended. It is not that they were not interesting, but I just failed to take nice notes on them.
– Prof. John Yin, Genetic and environmental impacts on the fitness of an RNA virus: computational models and wet-lab experiments. Systems Biology Theme Leader, Wisconsin Institute for Discovery, Department of Chemical and Biological Engineering, University of Wisconsin-Madison, USA (Homepage)
John Yin and his collaborators have developed a model with 140 equations that, given the genome of a single strand mRNA(+) virus, predicts the outcome of the infection. For example, a parameter of these equations is the order of the genes encoded. They have also experimentally verified how the efficiency of a virus infection changes when you shuffle the order of genes in the virus, and saw that if a certain gene is not in the first position the replication of the virus is not efficient.
– Pedro Beltrao: Evolution of phosphoregulation: from interactions to function. (Pablo’s blog on bioinformatics)
A study on the conservation of phosphorylation sites among eukaryotes. These sites tend to be loosely conserved, and the position of a phosphorylation site changes frequently. However, the total number of phosphorylation sites of a protein tends to be always the same in many organisms.
– Nicholas K. Priest: The role of compensatory mutation in the evolution of gene regulatory networks
They have simulated where, in a signaling network. a compensatory mutation would appear after a deleterious mutation. An example of compensatory mutation is the case of a insulin pathway where, after a deleterious mutation in one of the genes, the efficiency of an upstream gene increased. Personally, I wanted to ask some clarification on which definition of ‘compensatory mutation’ was used, because I felt that the term may be a bit vague.
– Laurence Loewe: Evolutionary systems biology and the distribution of mutational effects in the circadian clock of Ostreococcus
The circadian clock of the Ostreococcus organism is composed by a
pathway of genes that are activated or deactivated during daylight.
However, some mutations make so that the bacteria doesn’t follow
exactly this clock, and some genes are activated later or earlier than
expected. They are trying to put together a model to predict which
mutations and where would change the circadian clock and how.
Given the Genome (G) and Environment (E) where an organism develops,
EvolSysBio tries to predict the Phenotype (P).
– Prof. Juliette de Meaux: Asymmetric distribution of cis-regulatory differences reveal extant epigenetic differences between Arabidopsis genomes. Institute for Evolution and Biodiversity University of Münster, Germany (Homepage)
Plants are good models to study the effect of mutations in a organism, because they don’t move and can be exposed to the same environment. They have taken advantage of this to study plants with the same genotype and the differences in their phenotype produced by epigenetic.
– Chris Knight: Evolution of an environmental response network
They study the genes that regulates the “wrinkly” phenotype in yeast,
in response to a certain environment.
– David Alvarez-Ponze: Comparative genomics of the vertebrate insulin/TOR signal transduction pathway genes: A network-level analysis of selective pressures along the pathway
David analyzed the pathway of insulin signal transduction in vertebrates, and found a gradient of selection toward the genes in the downstream part. This is a nice work on the topic of whether the selective constraints which act on a gene are related to the position and degree of the gene in the network of interactions it is involved in. In our lab, we are working in the same field, and David is a friend of ours.
– Aurelien Mazurie: Evolution of metabolic network organization
They have developed the concept of NIP, which is a representation ofthe network of interactions within an organism. The NIP is the set of genes of an organism organized in KEGG pathways as if they were functional modules. So every organism is represented as a series of graphs composed by genes in the same KEGG pathway. Honestly I didn’t like very much this approach, because they used KEGG pathways as if they were kind of ‘functional modules’, while I believe they were not designed with this purpose and a cell is not aware of how his genes are annotated in KEGG (I should say that I am collaborating with Reactome which is a competitor to KEGG). Anyway, after calculating the network of reactions (NIP) of each organism, they calculate a vector of differences between each pair of species known, taking into consideration ~80 parameters like the connectivity of each pathway, etc.. Then they use a machine learning approach to identify the parameters that help more distinguish a group of organism from an another: for example, they discovered that networks in unicellular organisms tend to have higher betweeness centrality, while multicellular organisms have higher degrees. Their recent paper on ‘Evolution on Metabolic Network Organization‘ is interesting, especially the first figure which resumes differences on metabolic network organization among organisms. However, I would like to see the same pipeline applied to other annotations on pathway structure. I believe that their results are nice but the NIP step is not necessary.
Tamas Korcsmaros et al:, poster presentation: Comparison of signaling pathway proteins and the analysis of their evolutionary rate reveal the role of cross-talking proteins in three metazoans
I saw a very nice poster on database of signalling pathways, called signalink.org. It is a manually curated database on Signalling pathways in Human and a few other species. I talked a bit with one of the authors and he said that it almost all manually curated.