Genetic Analysis Workshop 15: gene expression analysis and approaches to detecting multiple functional loci

Mining functional gene modules linked with rheumatoid arthritis using a SNP-SNP network

The identification of functional gene modules that are derived from integration of information from different types of networks is a powerful strategy for interpreting the etiology of complex diseases such as rheumatoid arthritis (RA). Genetic variants are known to increase the risk of developing RA. Here, a novel method, the construction of a genetic network, was used to mine functional gene modules linked with RA. A polymorphism interaction analysis (PIA) algorithm was used to obtain cooperating single nucleotide polymorphisms (SNPs) that contribute to RA disease. The acquired SNP pairs were used to construct a SNP-SNP network. Sub-networks defined by hub SNPs were then extracted and turned into gene modules by mapping SNPs to genes using dbSNP database. We performed Gene Ontology (GO) analysis on each gene module, and some GO terms enriched in the gene modules can be used to investigate clustered gene function for better understanding RA pathogenesis. This method was applied to the Genetic Analysis Workshop 15 (GAW 15) RA dataset. The results show that genes involved in functional gene modules, such as CD160 (rs744877) and RUNX1 (rs2051179), are especially relevant to RA, which is supported by previous reports. Furthermore, the 43 SNPs involved in the identified gene modules were found to be the best classifiers when used as variables for sample classification.

Introduction to Genetic Analysis Workshop 15 summaries

The 15th biennial Genetic Analysis Workshop (GAW15) took place November 11-15, 2006 in St. Pete Beach, Florida. The workshop's primary focus was on the appropriate linkage, association, and other analyses of the increasingly large datasets generated by genetics research. A record number of participants (N=350) contributed 252 papers to GAW15. These contributions were organized into 17 presentation groups, with a range of 11 to 18 papers in each group (median of 15 papers per group). The data sets--or "problems"--for GAW15 included information from two real data sets and a simulated data set. The first problem utilizing real data included gene expression as the phenotype and genome-wide markers for linkage and association studies. The second problem allowed for detecting and characterizing genetic effects for rheumatoid arthritis. And the simulated problem was generated to reflect the data structure underlying the rheumatoid arthritis study. Further details on GAW15 are provided here, and the primary findings from the workshop are highlighted in the following group summary papers.

LASSO-Patternsearch algorithm with application to ophthalmology and genomic data

The LASSO-Patternsearch algorithm is proposed to efficiently identify patterns of multiple dichotomous risk factors for outcomes of interest in demographic and genomic studies. The patterns considered are those that arise naturally from the log linear expansion of the multivariate Bernoulli density. The method is designed for the case where there is a possibly very large number of candidate patterns but it is believed that only a relatively small number are important. A LASSO is used to greatly reduce the number of candidate patterns, using a novel computational algorithm that can handle an extremely large number of unknowns simultaneously. The patterns surviving the LASSO are further pruned in the framework of (parametric) generalized linear models. A novel tuning procedure based on the GACV for Bernoulli outcomes, modified to act as a model selector, is used at both steps. We applied the method to myopia data from the population-based Beaver Dam Eye Study, exposing physiologically interesting interacting risk factors. We then applied the the method to data from a generative model of Rheumatoid Arthritis based on Problem 3 from the Genetic Analysis Workshop 15, successfully demonstrating its potential to efficiently recover higher order patterns from attribute vectors of length typical of genomic studies.

Comparison of tagging single-nucleotide polymorphism methods in association analyses

Several methods to identify tagging single-nucleotide polymorphisms (SNPs) are in common use for genetic epidemiologic studies; however, there may be loss of information when using only a subset of SNPs. We sought to compare the ability of commonly used pairwise, multimarker, and haplotype-based tagging SNP selection methods to detect known associations with quantitative expression phenotypes. Using data from HapMap release 21 on unrelated Utah residents with ancestors from northern and western Europe (CEPH-Utah, CEU), we selected tagging SNPs in five chromosomal regions using ldSelect, Tagger, and TagSNPs. We found that SNP subsets did not substantially overlap, and that the use of trio data did not greatly impact SNP selection. We then tested associations between HapMap genotypes and expression phenotypes on 28 CEU individuals as part of Genetic Analysis Workshop 15. Relative to the use of all SNPs (n = 210 SNPs across all regions), most subset methods were able to detect single-SNP and haplotype associations. Generally, pairwise selection approaches worked extremely well, relative to use of all SNPs, with marked reductions in the number of SNPs required. Haplotype-based approaches, which had identified smaller SNP subsets, missed associations in some regions. We conclude that the optimal tagging SNP method depends on the true model of the genetic association (i.e., whether a SNP or haplotype is responsible); unfortunately, this is often unknown at the time of SNP selection. Additional evaluations using empirical and simulated data are needed.

Genome-wide association analyses of expression phenotypes

A number of issues arise when analyzing the large amount of data from high-throughput genotype and expression microarray experiments, including design and interpretation of genome-wide association studies of expression phenotypes. These issues were considered by contributions submitted to Group 1 of the Genetic Analysis Workshop 15 (GAW15), which focused on the association of quantitative expression data. These contributions evaluated diverse hypotheses, including those relevant to cancer and obesity research, and used various analytic techniques, many of which were derived from information theory. Several observations from these reports stand out. First, one needs to consider the genetic model of the trait of interest and carefully select which single nucleotide polymorphisms and individuals are included early in the design stage of a study. Second, by targeting specific pathways when analyzing genome-wide data, one can generate more interpretable results than agnostic approaches. Finally, for datasets with small sample sizes but a large number of features like the Genetic Analysis Workshop 15 dataset, machine learning approaches may be more practical than traditional parametric approaches.