Data mining and genetic algorithm based gene/SNP selection

Computational intelligence for genetic association study in complex diseases: review of theory and applications

“…Due to the complexity, there are relatively few works that are able to tackle this challenge. However, recently developed computational intelligence approaches for SNP-disease associations including genetic algorithms (Clark et al, 2005;Shah and Kusiak, 2004) neural networks (Ott, 2001;Motsinger et al, 2006aMotsinger et al, , 2006b, genetic programming (Moore and White, 2006a), evolutionary trees (Lam et al, 2000), evolutionary algorithms (Hubley et al, 2003) and various hybrid approaches, such as neural networks with genetic programming (Ritchie et al, 2003), genetic programming with multifactor dimensionality reduction (Moore and White, 2006b) and so on have demonstrated some promises, which we will summarise below.…”

“…Several surveys relating these approaches to disease mapping have been provided (Onkamo and Toivonen, 2006;Salem et al, 2005;Molitor et al, 2004;Shah and Kusiak, 2004;McKinney et al, 2006). For instance, Onkamo and Toivonen (2006) provided a survey of data mining approaches of disease mapping in bioinformatics (Onkamo and Toivonen, 2006); McKinney et al reviewed a number of different machine learning methods that have been applied to detect gene-gene interactions (McKinney et al, 2006).…”

Stages of Gene Regulatory Network Inference: the Evolutionary Algorithm Role

“…In this context, evolutionary techniques are known to cope well, as they benefit from mechanisms obtaining good solutions by searching a small portion only of the entire space, (Baeck et al, 2000). Consequently, there are several approaches that use EAs as wrapper methods for feature selection, for example Li (2001);Li et al (2004); Ooi & Tan (2003); Shah & Kusiak (2004); Souza & Carvalho (2005). One of these has also been applied to proteomics data, (Li, 2001;Li et al, 2004).…”

Genetic Programming — Introduction, Applications, Theory and Open Issues

“…Of particular medical interest are very wide data sets, with typically few samples and many inputs per sample. Recent examples include single nuclear polymorphisms (Shah and Kusiak 2004), chest pain (Bojarczuk et al 2000), and Affymetrix GeneChip microarray data (Langdon and Buxton 2004;Yu et al 2007). …”