A groupwise association test for rare mutations using a weighted sum statistic.

Madsen BE, Browning SR. · 2009

Read A groupwise association test for rare mutations using a weighted sum statistic. by Madsen BE, Browning SR. (2009) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc2633048.txt.

Introduction

Madsen and Browning proposed a statistical method in 2009 for detecting disease genes whose danger comes from many rare mutations rather than one common one. Standard genotyping studies are built for common variants and cannot tag rare mutations. Yet several rare faults in the same gene may together explain much disease, each too scarce to detect alone.

Their weighted-sum method tests a whole group of mutations jointly, weighting each by how much commoner it is in cases than controls. In simulations and in real ENCODE resequencing data it proved more powerful than alternative methods. The analysis shows a resequencing study of 1,000 to 7,000 people can find such a gene. Its mutations together may account for 2 percent of disease in the population. Each mutation alone accounts for far less.

Catalog record

CREATORMadsen BE, Browning SR.
YEAR2009
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC2633048

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Groupwise Association Test for Rare Mutations Using a Weighted Sum Statistic By Bo Eskerod Madsen, Sharon R. Browning, Nicholas J. Schork Published in: PLoS Genetics DOI: 10.1371/journal.pgen.1000384 ABSTRACT Resequencing is an emerging tool for identification of rare disease-associated mutations. Rare mutations are difficult to tag with SNP genotyping, as genotyping studies are designed to detect common variants. However, studies have shown that genetic heterogeneity is a probable scenario for common diseases, in which multiple rare mutations together explain a large proportion of the genetic basis for the disease. Thus, we propose a weighted-sum method to jointly analyse a group of mutations in order to test for groupwise association with disease status. For example, such a group of mutations may result from resequencing a gene. We compare the proposed weighted-sum method to alternative methods and show that it is powerful for identifying disease-associated genes, both on simulated and Encode data. Using the weighted-sum method, a resequencing study can identify a disease-associated gene with an overall population attributable risk (PAR) of 2%, even when each individual mutation has much lower PAR, using 1,000 to 7,000 affected and unaffected individuals, depending on the underlying genetic model. This study thus demonstrates that resequencing studies can identify important genetic associations, provided that specialised analysis methods, such as the weighted-sum method, are used. INTRODUCTION New technologies allow sequencing of parts of the genome of large groups of individuals [1], and hereby initiate the next generation of large scale association studies. Resequencing studies can directly identify millions of rare mutations in the genome, and may therefore be able to identify disease-mutations that are not tagged by panels of common SNPs [2]. Resequencing may thus hold the key to detecting associations in the presence of genetic heterogeneity, where the genetic component of disease-risk is determined by multiple rare mutations, each with a low marginal effect on disease-risk (i.e. low population attributable risk; PAR). Recent studies support the hypothesis that multiple rare mutations, each with a low marginal effect, may be a major player in genetic determination of susceptibility for some complex diseases [3]–[13]. Examples of genetically heterogeneous diseases include cystic fibrosis [14],[15], colorectal cancer [16] and probably schizophrenia [13]. Different genetic models may underlie genetic heterogeneity. One possibility is that multiple different variants located across the genome have independent influence on disease risk, such that each variant explains only a small fraction of all affected individuals. Another scenario is that the function of each haplotype of a gene is destroyed if one (or more) lethal mutations occur on the haplotype. In this manner, an individual must have at least one mutation on each of the two haplotypes to be predisposed for the disease (see the Recessive-Set model in Figure 1). In both of these models, the marginal PAR of each mutation may be very low, even when the disease is highly heritable. [Figure: Genetic models.Model descriptions and examples of predisposing genotypes are shown for the genetic models used. Lines symbolise haplotypes and dots symbolise disease-risk mutations.] Association studies using panels of common SNPs are well suited for identifying variants each with a relatively high PAR, whereas multiple rare variants, each with a small PAR, are difficult to identify using these methods [17]–[24]. In cases where a single (or very few) common variants are expected to be associated with a disease, a variant-by-variant approach using the strongest marginal signal for each tested variant may be beneficial (as discussed in [25] and [26]). On the other hand, when multiple rare mutations are expected to influence disease risk, an obvious approach is to group the variants according to function, such as genes, pathways and ultra conserved regions, and compare the group counts rather than the counts for each variant in the group. The rationale behind this grouping approach is that if many different mutations in a group affect disease risk, it may be beneficial to focus on the group rather than on each variant individually. The cohort allelic sums test (CAST) is an existing grouping method in which the number of individuals with one or more mutations in a group (e.g. gene) is compared between affected and unaffected individuals [5],[26],[27]. An alternative method using a grouping approach is the Combined Multivariate and Collapsing (CMC) method [26]. In this method all rare variants are collapsed, as in the CAST method, and the collapsed variants are treated as a single common variant which is analysed together with the other common variants using multivariate analysis [26]. In the CMC version used in [26], rare variants are defined as those having a minor allele frequency (MAF) of at most 1%. In this study, we focus on a scenario in which a group of multiple rare mutations has been identified. In functional regions, one may choose to include only probable disease susceptibility mutations (non-synonymous substitutions, frame shift mutations, etc) in the group of mutations. Using only probable disease susceptibility mutations has the benefit that random variation due to non-associated variants may decrease. In this manner, association studies of groups of rare probable disease-susceptibility variants may be able to identify genetically heterogeneous mutations, and hence complement genome-wide analysis of common SNPs. Grouping of mutations according to functional elements, such as genes, has the added advantage of focusing on causal relations between genes and diseases, rather than just identifying highly associated genomic regions. Furthermore, since many (millions of) mutations are expected to be identified in a resequencing study of thousands of individuals [28], grouping lowers the burden of multiple testing. We propose a weighted-sum method in which mutations are grouped according to function (e.g. gene), and each individual is scored by a weighted sum of the mutation counts. To test for an excess of mutations in affected individuals, we use permutation of disease status among affected and unaffected individuals. By using permutation, the method adjusts for the weighting of the mutations and the requirement that a mutation must be observed to be included in the study. Note that permutation of disease status results in correct type I error even in the presence of linkage disequilibrium (LD) [29],[30], although relatively low LD is expected between rare variants [26],[31],[32]. The weighted-sum method deviates from the CAST method [5],[27] by weighting the variants differently when determining the genetic load of an individual. By weighting the signals from each mutation, the weighted sum method accentuates mutations that are rare in the unaffected individuals, so that the test is not completely dominated by common mutations. In the CAST method, common variants will have a high impact on the group signal, and if many common mutations are present in a group, almost all individuals will have one or more mutations. To avoid this effect it may be necessary to use a threshold on the mutation-frequencies, as suggested in the CMC method [26]. A drawback of such frequency thresholds is that it can be difficult to select them in a biological meaningful way, and the outcome of the test will depend on the selection of thresholds. In the weighted-sum method we include mutations of all frequencies, but mutations are weighted according to their frequency in the unaffected individuals. METHODS Weighted-Sum Method The weighted-sum method compares the number of mutations in a group of variants

Read the complete text