question: I am trying to create an evolutionary algorithm utilizing epigenetic theory. Does anyone have information/insight/papers that might be of some use?

Hi Folks,

I am a compsci student, and I am planning to do my thesis based on evolutionary algorithms; specifically with epigenetics. I think that phenotype-genotype relationship in Evolutionary Algorithms can be improved if phenotypes did not translate literally to genotype. Exploring the solution space might be improved by utilizing epigenetic factors.

I have found quite a bit on epigenetic robotics and there is indeed a paper on ieee called the Epigenetic Algorithm, but I cannot access it. Most of the research on epigenetics seems to focus on using evolutionary algorithms to identify and study epigenetic factors rather than the other way around.

I would really appreciate if somebody could direct me to some papers, or methods, or explain to me what challenges lie ahead.

Cheers.

EDIT: Some update:

I have read the paper called Epigenetic Algorithm that was published on IEEE thanks to TleilaxuMaster. To be honest there are a few typos in the paper, then again I am a picky person. Apparently they were inspired in exactly the same fashion and took a through approach. But there is a lot room for new stuff. Here is the gist of it:

A comparison of EvolAlg vs Epigenetic Alg

and

A comparison of SwarmIntelligence vs Epigenetic Alg

• computational individuals represent genes/gene products rather than chromosomes, and they degrade over time rather than the effects of their behavior (pheromone trails in ant optimization).
• combinations of genes/gene products are evaluated for fitness rather than combinations of allele.
• they do not imitate their neighbours.
• each agent represents a specialized activity rather than a candidate solution, (some agents acting together might also represent a candidate solution).
• they use sequential and compositional storage mechanism (arrays that keep allele info and their relevance to other arrays, I suppose), as opposed to solely sequential storage mechanisms.
• unit of selection is an organization of individuals rather than a single fitting individual.
• each agent represents one gene, and a group of genes represent a solution as opposed to each agent representing a chromosome/genotype which may contain one or many genes
• while expression of a genotype changes the genotype remains the same as opposed to changing the genotype.
• agents are reproduced catalytically by other individuals rather than self producing.
• the behavior of agents is reproduced by influencing other agents rather than behavior being a product of direct spatial interaction.

Identified epigenetic Mechanisms:

• gene silencing: a result of histone modification, gene becomes inactive.
• paramutation: characteristic of a gene is remembered even if the gene is not present.
• bookmarking: transmit cellular memory of patterns of gene expression in a cell.
• genomic imprinting: certain gene are expressed in a parent of origin occurs from only one allele (one allele is enough for gene to manifest itself).
• position effect: the effect on the expression of the gene when the position in a chromosome is changed.
• reprogramming: remodeling of epigenetic markers (DNA methylation)
• transvection: interaction between corresponding allele of homologous chromosome which can lead to either gene activation or repression (did not get that one)
• maternal effect: genotype of mother expresses itself in the phenotype of offspring. (a dormant phenotype is expressed only in the offspring I suppose)
• x-inactivation: one of the two copies of the x-chromosome present in female mammals is inactivated.