Joshua Starmer

A multiple testing correction method for genetic association studies using correlated single nucleotide polymorphisms

Multiple testing is a challenging issue in genetic association studies using large numbers of single nucleotide polymorphism (SNP) markers, many of which exhibit linkage disequilibrium (LD). Failure to adjust for multiple testing appropriately may produce excessive false positives or overlook true positive signals. The Bonferroni method of adjusting for multiple comparisons is easy to compute, but is well known to be conservative in the presence of LD. On the other hand, permutation‐based corrections can correctly account for LD among SNPs, but are computationally intensive. In this work, we propose a new multiple testing correction method for association studies using SNP markers. We show that it is simple, fast and more accurate than the recently developed methods and is comparable to permutation‐based corrections using both simulated and real data. We also demonstrate how it might be used in whole‐genome association studies to control type I error. The efficiency and accuracy of the proposed method make it an attractive choice for multiple testing adjustment when there is high intermarker LD in the SNP data set. Genet. Epidemiol. 2008.

Avoiding the high Bonferroni penalty in genome-wide association studies

A major challenge in genome‐wide association studies (GWASs) is to derive the multiple testing threshold when hypothesis tests are conducted using a large number of single nucleotide polymorphisms. Permutation tests are considered the gold standard in multiple testing adjustment in genetic association studies. However, it is computationally intensive, especially for GWASs, and can be impractical if a large number of random shuffles are used to ensure accuracy. Many researchers have developed approximation algorithms to relieve the computing burden imposed by permutation. One particularly attractive alternative to permutation is to calculate the effective number of independent tests, Meff, which has been shown to be promising in genetic association studies. In this study, we compare recently developed Meff methods and validate them by the permutation test with 10,000 random shuffles using two real GWAS data sets: an Illumina 1M BeadChip and an Affymetrix GeneChip® Human Mapping 500K Array Set. Our results show that the simpleM method produces the best approximation of the permutation threshold, and it does so in the shortest amount of time. We also show that Meff is indeed valid on a genome‐wide scale in these data sets based on statistical theory and significance tests. The significance thresholds derived can provide practical guidelines for other studies using similar population samples and genotyping platforms. Genet. Epidemiol. 34:100–105, 2010.

Topoisomerases facilitate transcription of long genes linked to autism

Topoisomerases are expressed throughout the developing and adult brain and are mutated in some individuals with autism spectrum disorder (ASD). However, how topoisomerases are mechanistically connected to ASD is unknown. Here we find that topotecan, a topoisomerase 1 (TOP1) inhibitor, dose-dependently reduces the expression of extremely long genes in mouse and human neurons, including nearly all genes that are longer than 200 kilobases. Expression of long genes is also reduced after knockdown of Top1 or Top2b in neurons, highlighting that both enzymes are required for full expression of long genes. By mapping RNA polymerase II density genome-wide in neurons, we found that this length-dependent effect on gene expression was due to impaired transcription elongation. Interestingly, many high-confidence ASD candidate genes are exceptionally long and were reduced in expression after TOP1 inhibition. Our findings suggest that chemicals and genetic mutations that impair topoisomerases could commonly contribute to ASD and other neurodevelopmental disorders.

Site-Specific Silencing of Regulatory Elements as a Mechanism of X Inactivation

The inactive X chromosomes (Xi) physical territory is microscopically devoid of transcriptional hallmarks and enriched in silencing-associated modifications. How these microscopic signatures relate to specific Xi sequences is unknown. Therefore, we profiled Xi gene expression and chromatin states at high resolution via allele-specific sequencing in mouse trophoblast stem cells. Most notably, X-inactivated transcription start sites harbored distinct epigenetic signatures relative to surrounding Xi DNA. These sites displayed H3-lysine27-trimethylation enrichment and DNaseI hypersensitivity, similar to autosomal Polycomb targets, yet excluded Pol II and other transcriptional hallmarks, similar to nontranscribed genes. CTCF bound X-inactivated and escaping genes, irrespective of measured chromatin boundaries. Escape from X inactivation occurred within, and X inactivation was maintained exterior to, the area encompassed by Xist in cells subject to imprinted and random X inactivation. The data support a model whereby inactivation of specific regulatory elements, rather than a simple chromosome-wide separation from transcription machinery, governs gene silencing over the Xi.

Evidence of Xist RNA-independent initiation of mouse imprinted X-chromosome inactivation

XX female mammals undergo transcriptional silencing of most genes on one of their two X chromosomes to equalize X-linked gene dosage with XY males in a process referred to as X-chromosome inactivation (XCI). XCI is an example of epigenetic regulation. Once enacted in individual cells of the early female embryo, XCI is stably transmitted such that most descendant cells maintain silencing of that X chromosome. In eutherian mammals, XCI is thought to be triggered by the expression of the non-coding Xist RNA from the future inactive X chromosome (Xi); Xist RNA in turn is proposed to recruit protein complexes that bring about heterochromatinization of the Xi. Here we test whether imprinted XCI, which results in preferential inactivation of the paternal X chromosome (Xp), occurs in mouse embryos inheriting an Xp lacking Xist. We find that silencing of Xp-linked genes can initiate in the absence of paternal Xist; Xist is, however, required to stabilize silencing along the Xp. Xp-linked gene silencing associated with mouse imprinted XCI, therefore, can initiate in the embryo independently of Xist RNA.

Coexistent ARID1A–PIK3CA mutations promote ovarian clear-cell tumorigenesis through pro-tumorigenic inflammatory cytokine signalling

Ovarian clear-cell carcinoma (OCCC) is an aggressive form of ovarian cancer with high ARID1A mutation rates. Here we present a mutant mouse model of OCCC. We find that ARID1A inactivation is not sufficient for tumor formation, but requires concurrent activation of the phosphoinositide 3-kinase catalytic subunit, PIK3CA. Remarkably, the mice develop highly penetrant tumors with OCCC-like histopathology, culminating in hemorrhagic ascites and a median survival period of 7.5 weeks. Therapeutic treatment with the pan-PI3K inhibitor, BKM120, prolongs mouse survival by inhibiting tumor cell growth. Cross-species gene expression comparisons support a role for IL-6 inflammatory cytokine signaling in OCCC pathogenesis. We further show that ARID1A and PIK3CA mutations cooperate to promote tumor growth through sustained IL-6 overproduction. Our findings establish an epistatic relationship between SWI/SNF chromatin remodeling and PI3K pathway mutations in OCCC and demonstrate that these pathways converge on pro-tumorigenic cytokine signaling. We propose that ARID1A protects against inflammation-driven tumorigenesis.

A new model for random X chromosome inactivation

X chromosome inactivation (XCI) reduces the number of actively transcribed X chromosomes to one per diploid set of autosomes, allowing for dosage equality between the sexes. In eutherians, the inactive X chromosome in XX females is randomly selected. The mechanisms for determining both how many X chromosomes are present and which to inactivate are unknown. To understand these mechanisms, researchers have created X chromosome mutations and transgenes. Here, we introduce a new model of X chromosome inactivation that aims to account for the findings in recent studies, to promote a re-interpretation of existing data and to direct future experiments.

Human population structure detection via multilocus genotype clustering

BackgroundWe describe a hierarchical clustering algorithm for using Single Nucleotide Polymorphism (SNP) genetic data to assign individuals to populations. The method does not assume Hardy-Weinberg equilibrium and linkage equilibrium among loci in sample population individuals.ResultsWe show that the algorithm can assign sample individuals highly accurately to their corresponding ethnic groups in our tests using HapMap SNP data and it is also robust to admixed populations when tested with Perlegen SNP data. Moreover, it can detect fine-scale population structure as subtle as that between Chinese and Japanese by using genome-wide high-diversity SNP loci.ConclusionThe algorithm provides an alternative approach to the popular STRUCTURE program, especially for fine-scale population structure detection in genome-wide association studies. This is the first successful separation of Chinese and Japanese samples using random SNP loci with high statistical support.

KDM6 Demethylase Independent Loss of Histone H3 Lysine 27 Trimethylation during Early Embryonic Development

The early mammalian embryo utilizes histone H3 lysine 27 trimethylation (H3K27me3) to maintain essential developmental genes in a repressive chromatin state. As differentiation progresses, H3K27me3 is removed in a distinct fashion to activate lineage specific patterns of developmental gene expression. These rapid changes in early embryonic chromatin environment are thought to be dependent on H3K27 demethylases. We have taken a mouse genetics approach to remove activity of both H3K27 demethylases of the Kdm6 gene family, Utx (Kdm6a, X-linked gene) and Jmjd3 (Kdm6b, autosomal gene). Male embryos null for active H3K27 demethylation by the Kdm6 gene family survive to term. At mid-gestation, embryos demonstrate proper patterning and activation of Hox genes. These male embryos retain the Y-chromosome UTX homolog, UTY, which cannot demethylate H3K27me3 due to mutations in catalytic site of the Jumonji-C domain. Embryonic stem (ES) cells lacking all enzymatic KDM6 demethylation exhibit a typical decrease in global H3K27me3 levels with differentiation. Retinoic acid differentiations of these ES cells demonstrate loss of H3K27me3 and gain of H3K4me3 to Hox promoters and other transcription factors, and induce expression similar to control cells. A small subset of genes exhibit decreased expression associated with reduction of promoter H3K4me3 and some low-level accumulation of H3K27me3. Finally, Utx and Jmjd3 mutant mouse embryonic fibroblasts (MEFs) demonstrate dramatic loss of H3K27me3 from promoters of several Hox genes and transcription factors. Our results indicate that early embryonic H3K27me3 repression can be alleviated in the absence of active demethylation by the Kdm6 gene family.

AWclust: point-and-click software for non-parametric population structure analysis

BackgroundPopulation structure analysis is important to genetic association studies and evolutionary investigations. Parametric approaches, e.g. STRUCTURE and L-POP, usually assume Hardy-Weinberg equilibrium (HWE) and linkage equilibrium among loci in sample population individuals. However, the assumptions may not hold and allele frequency estimation may not be accurate in some data sets. The improved version of STRUCTURE (version 2.1) can incorporate linkage information among loci but is still sensitive to high background linkage disequilibrium. Nowadays, large-scale single nucleotide polymorphisms (SNPs) are becoming popular in genetic studies. Therefore, it is imperative to have software that makes full use of these genetic data to generate inference even when model assumptions do not hold or allele frequency estimation suffers from high variation.ResultsWe have developed point-and-click software for non-parametric population structure analysis distributed as an R package. The software takes advantage of the large number of SNPs available to categorize individuals into ethnically similar clusters and it does not require assumptions about population models. Nor does it estimate allele frequencies. Moreover, this software can also infer the optimal number of populations.ConclusionOur software tool employs non-parametric approaches to assign individuals to clusters using SNPs. It provides efficient computation and an intuitive way for researchers to explore ethnic relationships among individuals. It can be complementary to parametric approaches in population structure analysis.