Composite Linkage Mapping Using Back-Propagation Artificial Neural Networks
With the completion of the HapMap project, largescale, high-density single-nucleotide polymorphism (SNP) markers and information on haplotype structure and frequencies become available. Although meta-analysis could combine the data by placing markers on a common genetic map and performing linkage analysis on the whole sample jointly, how to adjust the map distance of those markers with no crossover data sets in available databases is still an important problem. In this paper, back-propagation artificial neural network was be used to combine the linkage mapping. The results showed composite genetic map can be got after adjusted the small sample map distance curve with a criteria curve, a large sample size curve or a map got from meta-analysis, and the composite genetic map with more precise and dense markers could also be got using an appropriate networks trained with crossover marker data. The small genetic maps could also be spliced by using an appropriate networks trained with crossover marker data, the network structure with different hidden neurons was also very important for splicing the genetic maps from different literatures. The meta-analysis combining appropriate networks trained with crossover marker data was a good selection for composite mapping.
X. B. Li and X. L. Yu, "Composite Linkage Mapping Using Back-Propagation Artificial Neural Networks", Advanced Materials Research, Vols. 108-111, pp. 285-290, 2010