6533b7cefe1ef96bd1256f72

RESEARCH PRODUCT

false

subject

description

Genome-Wide-Association-Studies have become a powerful method to link point mutations (e.g. single nucleotide polymorphisms (SNPs)) to a certain phenotype or a disease. However, their power to detect SNPs associated to polygenic diseases such as Alzheimer's Disease (AD) is limited, since they can only infer the pairwise relation of single SNPs to the phenotype and ignore possible effects of various SNP combinations. The common method to probe these possible complex genetic patterns is to compute a measure called linkage disequilibrium (LD). Despite the fact that several predictive patterns found with LD could successfully be applied to medical diagnosis, this measure still holds several drawbacks as for example the difficulty to confirm and replicate experimental results as well as its sensitivity to statistical biases. Here, we present the application of an alternative method, Linkage Probability (LP) for genetic pattern identification that provides the posterior probability of a relation between two categorical data sets and simultaneously considers potential biases from latent variables, such as the recombination rate or the genetic structure of a population. By applying the LP framework to data from the ADSP-Project, we show that changes of linkage patterns between SNPs can be associated to Alzheimer's disease. Common genomic relation measures still fail to extract this link.