Jurg Ott

Because current molecular haplotyping methods are expensive and not amenable to automation, many researchers rely on statistical methods to infer haplotype pairs from multilocus genotypes, and subsequently treat these inferred haplotype pairs as observations. These procedures are prone to haplotype misclassification. We examine the effect of these misclassification errors on the false-positive rate and power for two association tests. These tests include the standard likelihood ratio test (LRT_std) and a likelihood ratio test that employs a double-sampling approach to allow for the misclassification inherent in the haplotype inference procedure (LRT_ae). We aim to determine the cost–benefit relationship of increasing the proportion of individuals with molecular haplotype measurements in addition to genotypes to raise the power gain of the LRT_ae over the LRT_std. This analysis should provide a guideline for determining the minimum number of molecular haplotypes required for desired power. Our simulations under the null hypothesis of equal haplotype frequencies in cases and controls indicate that (1) for each statistic, permutation methods maintain the correct type I error; (2) specific multilocus genotypes that are misclassified as the incorrect haplotype pair are consistently misclassified throughout each entire dataset; and (3) our simulations under the alternative hypothesis showed a significant power gain for the LRT_ae over the LRT_std for a subset of the parameter settings. Permutation methods should be used exclusively to determine significance for each statistic. For fixed cost, the power gain of the LRT_ae over the LRT_std varied depending on the relative costs of genotyping, molecular haplotyping, and phenotyping. The LRT_ae showed the greatest benefit over the LRT_std when the cost of phenotyping was very high relative to the cost of genotyping. This situation is likely to occur in a replication study as opposed to a whole-genome association study. Localizing genes for complex genetic diseases presents a major challenge. Recent technological advances such as genotyping arrays containing hundreds of thousands of genomic “landmarks,” and databases cataloging these “landmarks” and the levels of correlation between them, have aided in these endeavors. To utilize these resources most effectively, many researchers employ a gene-mapping technique called haplotype-based association in order to examine the variation present at multiple genomic sites jointly for a role in and/or an association with the disease state. Although methods that determine haplotype pairs directly by biological assays are currently available, they rarely are used due to their expense and incongruity to automation. Statistical methods provide an inexpensive, relatively accurate means to determine haplotype pairs. However, these statistical methods can provide erroneous results. In this article, the authors compare a standard statistical method for performing a haplotype-based association test with a method that accounts for the misclassification of haplotype pairs as part of the test. Under a number of feasible scenarios, the performance of the new test exceeded that of the standard test.

With the increasing amount of data generated in molecular genetics laboratories, it is often difficult to make sense of results because of the vast number of different outcomes or variables studied. Examples include expression levels for large numbers of genes and haplotypes at large numbers of loci. It is then natural to group observations into smaller numbers of classes that allow for an easier overview and interpretation of the data. This grouping is often carried out in multiple steps with the aid of hierarchical cluster analysis, each step leading to a smaller number of classes by combining similar observations or classes. At each step, either implicitly or explicitly, researchers tend to interpret results and eventually focus on that set of classes providing the "best" (most significant) result. While this approach makes sense, the overall statistical significance of the experiment must include the clustering process, which modifies the grouping structure of the data and often removes variation. For hierarchically clustered data, we propose considering the strongest result or, equivalently, the smallest p-value as the experiment-wise statistic of interest and evaluating its significance level for a global assessment of statistical significance. We apply our approach to datasets from haplotype association and microarray expression studies where hierarchical clustering has been used. In all of the cases we examine, we find that relying on one set of classes in the course of clustering leads to significance levels that are too small when compared with the significance level associated with an overall statistic that incorporates the process of clustering. In other words, relying on one step of clustering may furnish a formally significant result while the overall experiment is not significant.

Utilizing DNA samples from 91 Afrikaner nuclear families with one or more affected children, five genomic regions on chromosomes 2p, 8q, 11q, 20q, and 21q that gave evidence for association with GTS in previous case‐control association studies were investigated for linkage and association with GTS. Highly polymorphic markers with mean heterozygosity of 0.77 were typed and resulting genotypes evaluated using single marker transmission disequilibrium (TDT), single marker haplotype relative risk (HRR), and multi‐marker “extended” TDT and HRR methods. Single marker TDT analysis showed evidence for linkage or association, with p‐values near 0.05, for markers D2S139, GATA28F12, and D11S1377 on chromosomes 2p11, 8q22 and 11q23‐24, respectively. Extended, two‐locus TDT and HRR analysis provided further evidence for linkage or association on chromosome 2 with p‐values of 0.007 and 0.025, and chromosome 8 with p‐values of 0.059 and 0.013, respectively. These results provide important additional evidence for the location of GTS susceptibility loci.

After a brief introduction to neural networks, specific applications are proposed for the analysis of association between single nucleotide polymorphisms and putative disease genes in case-control studies.

For large numbers of marker loci in a genomic scan for disease loci, we propose a novel 2‐stage approach for linkage or association analysis. The two stages are (1) selection of a subset of markers that are ‘important’ for the trait studied, and (2) modelling interactions among markers and between markers and trait. Here we focus on stage 1 and develop a selection method based on a 2‐level nested bootstrap procedure. The method is applied to single nucleotide polymorphisms (SNPs) data in a cohort study of heart disease patients. Out of the 89 original SNPs the method selects 11 markers as being ‘important’. Conventional backward stepwise logistic regression on the 89 SNPs selects 7 markers, which are a subset of the 11 markers chosen by our method.

Scan statistics are applied to combine information on multiple contiguous genetic markers used in a genome screen for susceptibility loci. This information may be, for example, allele sharing proportions for sib pairs or logarithm of odds (lod) scores in general small families. We focus on a dichotomous outcome variable, for example, case and control individuals or affected-affected versus affected-unaffected siblings, and suitable single-marker statistics. A significant scan statistic based on the single-marker statistics represents evidence of the presence of a susceptibility gene. For a given length of the scan statistic, we assess its significance by Monte Carlo permutation tests. Comparing P values for varying lengths of scan statistics, we treat the smallest observed P value as our statistic of interest and determine its overall significance level. We applied this method to a genome screen with autism families. The result was informative and surprising: A susceptibility region was found (genome-wide significance level, P = 0.038), which is missed with conventional approaches.