Complex Systems and Artificial Cognitive Processes in Plant Genetics

Linear models and distributive assumptions are the basis of the traditional dissection of complex phenotypes in Genetics: a sound example is the Fisher’s infinitesimal model in the analysis and interpretation of quantitative traits. This simple but effective approach is based on a black box strategy in which unspecified natural functions associate the input (unknown) variables, the genes, to the output ones, the phenotypes. Nowadays the increasing knowledge on genomics, gene expression and metabolic networks strongly suggests the opportunity to open the black box and to fill it of sound interpretation of the relevant paths between genes and phenotype or its components. The class of artificial cognitive processes can be an appealing solution. More particularly, some procedures, based on computer simulations and Montecarlo methods, are available. They reach solutions by means of an iterative self-learning process based on the principles, for instance, of natural selection and evolution.