Data Mining Applied to Linkage Disequilibrium Mapping

Abstract

We introduce a new method for linkage disequilibrium mapping: haplotype pattern mining (HPM). The method, inspired by data mining methods, is based on discovery of recurrent patterns. We define a class of useful haplotype patterns in genetic case-control data and use the algorithm for finding disease-associated haplotypes. The haplotypes are ordered by their strength of association with the phenotype, and all haplotypes exceeding a given threshold level are used for prediction of disease susceptibility-gene location. The method is model-free, in the sense that it does not require (and is unable to utilize) any assumptions about the inheritance model of the disease. The statistical model is nonparametric. The haplotypes are allowed to contain gaps, which improves the method's robustness to mutations and to missing and erroneous data. Experimental studies with simulated microsatellite and SNP data show that the method has good localization power in data sets with large degrees of phenocopies and with lots of missing and erroneous data. The power of HPM is roughly identical for marker maps at a density of 3 single-nucleotide polymorphisms/cM or 1 microsatellite/cM. The capacity to handle high proportions of phenocopies makes the method promising for complex disease mapping. An example of correct disease susceptibility-gene localization with HPM is given with real marker data from families from the United Kingdom affected by type 1 diabetes. The method is extendable to include environmental covariates or phenotype measurements or to find several genes simultaneously.

A Second-Generation Association Study of the 5q31 Cytokine Gene Cluster and the Interleukin-4 Receptor in Asthma

“…For haplotype analysis we used the HPM method, which is powerful in locating a disease-causing gene when one is known to exist [22]. By allowing gaps in haplotype patterns, it is also more robust for genotyping errors, marker mutations, unrecognized recombinations, and missing data than the haplotype analysis we used in our first study [18,22].…”

“…By allowing gaps in haplotype patterns, it is also more robust for genotyping errors, marker mutations, unrecognized recombinations, and missing data than the haplotype analysis we used in our first study [18,22]. Also, the HPM parameters can be optimized for each data set by studying individual features of genotype data such as marker information, marker density, and missing genotypes.…”

Advances in statistical human genetics over the last 25 years

“…The overall time complexity of the algorithm is O (M N 2 ), where M is the total number of marker loci and N is the sample size, which is around hundreds in many real data sets. Preliminary experimental results on the simulated data given in Reference [185], and on an HLA real data set [190] with known disease gene location for type 1 diabetes, show that this method can predict gene location with high accuracy and its performance increases with denser markers. Phase information can also be obtained from the genotypes of family members, though this does not usually completely resolve phase ambiguity [48,52,191,192].…”

“…The effect of violations of these assumptions is unpredictable in general. A non-parametric method called HPM (haplotype pattern mining) [185], inspired by data mining methods, has been proposed to identify disease associated haplotype patterns from case-control data. HPM does not require any assumptions about the inheritance pattern and has good localization power, even when the number of phenocopies, i.e.…”

“…Third, the results obtained by using a permutation test for HPM to predict gene locations can be inconclusive and the prediction accuracy may be worse with dense (e.g. SNP) markers [185], which is undesirable and greatly limits the utility of the method.…”