Use of an artificial neural network to detect association between a disease and multiple marker genotypes

Single nucleotide polymorphisms (SNPs) are very common throughout the genom e and hence are potentially valuable for mapping disease susceptibility loc i by detecting association between SNP markers and disease. However as SNPs are biallelic they may have relatively little power in association studies compared with the information that would be obtainable if marker haplotype s were available and could be used efficiently. Modelling the evolutionary events leading to linkage disequilibrium is very complex and many methods t hat seek to use information from multiple markers simultaneously need to ma ke simplifying assumption sand may only be applicable when marker haplotype s, rather than genotypes, are available for analysis. We explore the proper ties of a simple application of a standard artificial neural network to thi s problem. The pattern-recognition properties of the network are used in th e hope that marker haplotypes implicit in the genotypes will differ between cases and controls in a was which will lead to the network being able to c lassify the subjects correctly, according to their marker genotype. This me thod makes no assumption sat all regarding population history or the marker map, and can be applied to genotypes, as would be available from a simple case-control sample, without any need to determine haplotypes. Through appl ication to data simulated under a very wide range of assumptions we show th at such an analysis produces a useful augmentation in power above that whic h would be achieved by testing each marker individually, in particular when more than one mutation has occurred in a disease gene at different points in evolution. The application of neural networks to such problems shows con siderable promise and further work could usefully be directed towards optim ising the design and implementation of such networks.