Evolving fuzzy rules to model gene expression

Abstract

This paper develops an algorithm that extracts explanatory rules from microarray data, which we treat as time series, using genetic programming (GP) and fuzzy logic. Reverse polish notation is used (RPN) to describe the rules and to facilitate the GP approach. The algorithm also allows for the insertion of prior knowledge, making it possible to find sets of rules that include the relationships between genes already known. The algorithm proposed is applied to problems arising in the construction of gene regulatory networks, using two different sets of real data from biological experiments on the Arabidopsis thaliana cold response and the rat central nervous system, respectively. The results show that the proposed technique can fit data to a pre-defined precision even in situations where the data set has thousands of features but only a limited number of points in time are available, a situation in which traditional statistical alternatives encounter difficulties, due to the scarcity of time points.

Stages of Gene Regulatory Network Inference: the Evolutionary Algorithm Role

“…Analysis of this kind of model is very complex, so viewing the network 'top-down', in order to be able to analyse it globally, is the aim of coarse-grained models, ( e.g. Linden & Bhaya (2007); Maki et al (2001); Repsilber et al (2002)). Other authors, (e.g.…”

“…This approach has the advantage of being more intuitive, as relationships between genes are expressed using natural language. One such model uses fuzzy rules, (Linden & Bhaya, 2007), which are based on the notion of fuzzy sets. These sets have imprecise boundaries, defined by a membership function: applied to any element in the universe, they return a number in the interval [0,1], representing the degree to which that element is a member of the current set.…”

“…Linden & Bhaya (2007) introduced a method of inferring fuzzy rules from microarray data, using genetic programming. The algorithm uses the reverse Polish notation 6 for rules, which can be easily represented as trees, with three Boolean operators for the conditions: NOT, AND and OR.…”

The evolution of higher-level biochemical reaction models

“…Stochastic oscillating behaviors are evolved using GP in Leier et al (2006), Imada and Ross (2010) and Ross (2010). Fuzzy rules are evolved with GP in Linden and Bhaya (2007), which although not probabilistic, does account for stochastic behavior. The problem of evolving stochastic bio-networks is closely related to the more general problem of evolving models for noisy, chaotic time series.…”

A Gene Regulatory Network Prediction Method using Particle Swarm Optimization and Genetic Algorithm

“…The GRN of a living organism can be inferred from its microarray dataset. Many methods have been proposed for predicting GRNs based on different techniques such as Boolean Networks [2], Perti Nets [3], Dynamic Bayesian Networks [4], and we also have other approaches which are based on soft computing approaches such as genetic programming [5], Fuzzy Rules [6].…”