Multiple Testing With an Empirical Alternative Hypothesis

An optimal multiple testing procedure is identified for linear hypotheses under the general linear model, maximizing the expected number of false null hypotheses rejected at any significance level. The optimal procedure depends on the unknown data-generating distribution, but can be consistently estimated. Drawing information together across many hypotheses, the estimated optimal procedure provides an empirical alternative hypothesis by adapting to underlying patterns of departure from the null. Proposed multiple testing procedures based on the empirical alternative are evaluated through simulations and an application to gene expression microarray data. Compared to a standard multiple testing procedure, it is not unusual for use of an empirical alternative hypothesis to increase by 50% or more the number of true positives identified at a given significance level.

Comparison of multiple comparison methods for identifying differential gene expression in simulated and real papillary thyroid cancer microarray data

The difficulty of detecting differential gene expression in microarray data has existed for many years. Several correction procedures try to avoid the family-wise error rate in multiple comparison process, including the Bonferroni and Sidak single-step p-value adjustments, Holm's step-down correction method, and Benjamini and Hochberg's false discovery rate (FDR) correction procedure. Each multiple comparison technique has its advantages and weaknesses. We studied each multiple comparison method through numerical studies (simulations) and applied the methods to the real exploratory DNA microarray data, which detect of molecular signatures in papillary thyroid cancer (PTC) patients. According to our results of simulation studies, Benjamini and Hochberg step-up FDR controlling procedure is the best process among these multiple comparison methods and we discovered 1277 potential biomarkers among 54675 probe sets after applying the Benjamini and Hochberg's method to PTC microarray data.^

Longitudinal and logistic analysis of endocrine hormone profiles, physical maturation, and candidate genes related to pubertal events in children

Children who experience early pubertal development have an increased risk of developing cancer (breast, ovarian, and testicular), osteoporosis, insulin resistance, and obesity as adults. Early pubertal development has been associated with depression, aggressiveness, and increased sexual prowess. Possible explanations for the decline in age of pubertal onset include genetics, exposure to environmental toxins, better nutrition, and a reduction in childhood infections. In this study we (1) evaluated the association between 415 single nucleotide polymorphisms (SNPs) from hormonal pathways and early puberty, defined as menarche prior to age 12 in females and Tanner Stage 2 development prior to age 11 in males, and (2) measured endocrine hormone trajectories (estradiol, testosterone, and DHEAS) in relation to age, race, and Tanner Stage in a cohort of children from Project HeartBeat! At the end of the 4-year study, 193 females had onset of menarche and 121 males had pubertal staging at age 11. African American females had a younger mean age at menarche than Non-Hispanic White females. African American females and males had a lower mean age at each pubertal stage (1-5) than Non-Hispanic White females and males. African American females had higher mean BMI measures at each pubertal stage than Non-Hispanic White females. Of the 415 SNPs evaluated in females, 22 SNPs were associated with early menarche, when adjusted for race ( p<0.05), but none remained significant after adjusting for multiple testing by False Discovery Rate (p<0.00017). In males, 17 SNPs were associated with early pubertal development when adjusted for race (p<0.05), but none remained significant when adjusted for multiple testing (p<0.00017). ^ There were 4955 hormone measurements taken during the 4-year study period from 632 African American and Non-Hispanic White males and females. On average, African American females started and ended the pubertal process at a younger age than Non-Hispanic White females. The mean age of Tanner Stage 2 breast development in African American and Non-Hispanic White females was 9.7 (S.D.=0.8) and 10.2 (S.D.=1.1) years, respectively. There was a significant difference by race in mean age for each pubertal stage, except Tanner Stage 1 for pubic hair development. Both Estradiol and DHEAS levels in females varied significantly with age, but not by race. Estradiol and DHEAS levels increased from Tanner Stage 1 to Tanner Stage 5.^ African American males had a lower mean age at each Tanner Stage of development than Non-Hispanic White males. The mean age of Tanner Stage 2 genital development in African American and Non-Hispanic White males was 10.5 (S.D.=1.1) and 10.8 (S.D.=1.1) years, respectively, but this difference was not significant (p=0.11). Testosterone levels varied significantly with age and race. Non-Hispanic White males had higher levels of testosterone than African American males from Tanner Stage 1-4. Testosterone levels increased for both races from Tanner Stage 1 to Tanner Stage 5. Testosterone levels had the steepest increase from ages 11-15 for both races. DHEAS levels in males varied significantly with age, but not by race. DHEAS levels had the steepest increase from ages 14-17. ^ In conclusion, African American males and females experience pubertal onset at a younger age than Non-Hispanic White males and females, but in this study, we could not find a specific gene that explained the observed variation in age of pubertal onset. Future studies with larger study populations may provide a better understanding of the contribution of genes in early pubertal onset.^

RNA-SEQUENCING APPLICATIONS: GENE EXPRESSION QUANTIFICATION AND METHYLATOR PHENOTYPE IDENTIFICATION

My dissertation focuses on two aspects of RNA sequencing technology. The first is the methodology for modeling the overdispersion inherent in RNA-seq data for differential expression analysis. This aspect is addressed in three sections. The second aspect is the application of RNA-seq data to identify the CpG island methylator phenotype (CIMP) by integrating datasets of mRNA expression level and DNA methylation status. Section 1: The cost of DNA sequencing has reduced dramatically in the past decade. Consequently, genomic research increasingly depends on sequencing technology. However it remains elusive how the sequencing capacity influences the accuracy of mRNA expression measurement. We observe that accuracy improves along with the increasing sequencing depth. To model the overdispersion, we use the beta-binomial distribution with a new parameter indicating the dependency between overdispersion and sequencing depth. Our modified beta-binomial model performs better than the binomial or the pure beta-binomial model with a lower false discovery rate. Section 2: Although a number of methods have been proposed in order to accurately analyze differential RNA expression on the gene level, modeling on the base pair level is required. Here, we find that the overdispersion rate decreases as the sequencing depth increases on the base pair level. Also, we propose four models and compare them with each other. As expected, our beta binomial model with a dynamic overdispersion rate is shown to be superior. Section 3: We investigate biases in RNA-seq by exploring the measurement of the external control, spike-in RNA. This study is based on two datasets with spike-in controls obtained from a recent study. We observe an undiscovered bias in the measurement of the spike-in transcripts that arises from the influence of the sample transcripts in RNA-seq. Also, we find that this influence is related to the local sequence of the random hexamer that is used in priming. We suggest a model of the inequality between samples and to correct this type of bias. Section 4: The expression of a gene can be turned off when its promoter is highly methylated. Several studies have reported that a clear threshold effect exists in gene silencing that is mediated by DNA methylation. It is reasonable to assume the thresholds are specific for each gene. It is also intriguing to investigate genes that are largely controlled by DNA methylation. These genes are called “L-shaped” genes. We develop a method to determine the DNA methylation threshold and identify a new CIMP of BRCA. In conclusion, we provide a detailed understanding of the relationship between the overdispersion rate and sequencing depth. And we reveal a new bias in RNA-seq and provide a detailed understanding of the relationship between this new bias and the local sequence. Also we develop a powerful method to dichotomize methylation status and consequently we identify a new CIMP of breast cancer with a distinct classification of molecular characteristics and clinical features.

Significance analysis of microarrays applied to the ionizing radiation response

Microarrays can measure the expression of thousands of genes to identify changes in expression between different biological states. Methods are needed to determine the significance of these changes while accounting for the enormous number of genes. We describe a method, Significance Analysis of Microarrays (SAM), that assigns a score to each gene on the basis of change in gene expression relative to the standard deviation of repeated measurements. For genes with scores greater than an adjustable threshold, SAM uses permutations of the repeated measurements to estimate the percentage of genes identified by chance, the false discovery rate (FDR). When the transcriptional response of human cells to ionizing radiation was measured by microarrays, SAM identified 34 genes that changed at least 1.5-fold with an estimated FDR of 12%, compared with FDRs of 60 and 84% by using conventional methods of analysis. Of the 34 genes, 19 were involved in cell cycle regulation and 3 in apoptosis. Surprisingly, four nucleotide excision repair genes were induced, suggesting that this repair pathway for UV-damaged DNA might play a previously unrecognized role in repairing DNA damaged by ionizing radiation.

Methylomic and Transcriptomic Insights Linking Air Pollution and Atherosclerosis

Thesis (Ph.D.)--University of Washington, 2016-06

Application of mixture models to detect differentially expressed genes

An important and common problem in microarray experiments is the detection of genes that are differentially expressed in a given number of classes. As this problem concerns the selection of significant genes from a large pool of candidate genes, it needs to be carried out within the framework of multiple hypothesis testing. In this paper, we focus on the use of mixture models to handle the multiplicity issue. With this approach, a measure of the local FDR (false discovery rate) is provided for each gene. An attractive feature of the mixture model approach is that it provides a framework for the estimation of the prior probability that a gene is not differentially expressed, and this probability can subsequently be used in forming a decision rule. The rule can also be formed to take the false negative rate into account. We apply this approach to a well-known publicly available data set on breast cancer, and discuss our findings with reference to other approaches.

Using mixture models to detect differentially expressed genes

An important and common problem in microarray experiments is the detection of genes that are differentially expressed in a given number of classes. As this problem concerns the selection of significant genes from a large pool of candidate genes, it needs to be carried out within the framework of multiple hypothesis testing. In this paper, we focus on the use of mixture models to handle the multiplicity issue. With this approach, a measure of the local false discovery rate is provided for each gene, and it can be implemented so that the implied global false discovery rate is bounded as with the Benjamini-Hochberg methodology based on tail areas. The latter procedure is too conservative, unless it is modified according to the prior probability that a gene is not differentially expressed. An attractive feature of the mixture model approach is that it provides a framework for the estimation of this probability and its subsequent use in forming a decision rule. The rule can also be formed to take the false negative rate into account.

A simple implementation of a normal mixture approach to differential gene expression in multiclass microarrays

Motivation: An important problem in microarray experiments is the detection of genes that are differentially expressed in a given number of classes. We provide a straightforward and easily implemented method for estimating the posterior probability that an individual gene is null. The problem can be expressed in a two-component mixture framework, using an empirical Bayes approach. Current methods of implementing this approach either have some limitations due to the minimal assumptions made or with more specific assumptions are computationally intensive. Results: By converting to a z-score the value of the test statistic used to test the significance of each gene, we propose a simple two-component normal mixture that models adequately the distribution of this score. The usefulness of our approach is demonstrated on three real datasets.

Mixture models for detecting differentially expressed genes in microarrays

An important and common problem in microarray experiments is the detection of genes that are differentially expressed in a given number of classes. As this problem concerns the selection of significant genes from a large pool of candidate genes, it needs to be carried out within the framework of multiple hypothesis testing. In this paper, we focus on the use of mixture models to handle the multiplicity issue. With this approach, a measure of the local FDR (false discovery rate) is provided for each gene. An attractive feature of the mixture model approach is that it provides a framework for the estimation of the prior probability that a gene is not differentially expressed, and this probability can subsequently be used in forming a decision rule. The rule can also be formed to take the false negative rate into account. We apply this approach to a well-known publicly available data set on breast cancer, and discuss our findings with reference to other approaches.

Changes in Brain Resting-state Functional Connectivity Associated with Peripheral Nerve Block: A Pilot Study.

BACKGROUND: Limited information exists on the effects of temporary functional deafferentation (TFD) on brain activity after peripheral nerve block (PNB) in healthy humans. Increasingly, resting-state functional connectivity (RSFC) is being used to study brain activity and organization. The purpose of this study was to test the hypothesis that TFD through PNB will influence changes in RSFC plasticity in central sensorimotor functional brain networks in healthy human participants. METHODS: The authors achieved TFD using a supraclavicular PNB model with 10 healthy human participants undergoing functional connectivity magnetic resonance imaging before PNB, during active PNB, and during PNB recovery. RSFC differences among study conditions were determined by multiple-comparison-corrected (false discovery rate-corrected P value less than 0.05) random-effects, between-condition, and seed-to-voxel analyses using the left and right manual motor regions. RESULTS: The results of this pilot study demonstrated disruption of interhemispheric left-to-right manual motor region RSFC (e.g., mean Fisher-transformed z [effect size] at pre-PNB 1.05 vs. 0.55 during PNB) but preservation of intrahemispheric RSFC of these regions during PNB. Additionally, there was increased RSFC between the left motor region of interest (PNB-affected area) and bilateral higher order visual cortex regions after clinical PNB resolution (e.g., Fisher z between left motor region of interest and right and left lingual gyrus regions during PNB, -0.1 and -0.6 vs. 0.22 and 0.18 after PNB resolution, respectively). CONCLUSIONS: This pilot study provides evidence that PNB has features consistent with other models of deafferentation, making it a potentially useful approach to investigate brain plasticity. The findings provide insight into RSFC of sensorimotor functional brain networks during PNB and PNB recovery and support modulation of the sensory-motor integration feedback loop as a mechanism for explaining the behavioral correlates of peripherally induced TFD through PNB.

The differential impact of oxytocin receptor (OXTR) genotypes on the risk of autism spectrum disorder (ASD) and resulting social communication deficits

Background: Autism spectrum disorder (ASD) is multifactorial and is likely the result of complex interactions between multiple environmental and genetic factors. Recently, it has been suggested that each symptom cluster of the disorder, such as poor social communication, may be mediated by different genetic influences. Genes in the oxytocin pathway, which mediates social behaviours in humans, have been studied with single nucleotide polymorphisms (SNPs) in the oxytocin receptor gene (OXTR) being implicated in ASD. This thesis examines the presence of different oxytocin receptor genotypes, and their associations with ASD and resulting social communication deficits. Methods: The relationship between four OXTR variants and ASD was evaluated in 607 ASD simplex (SPX) families. Cases were compared to their unaffected siblings using a conditional logistic approach. Odds ratios and associated 95 percent confidence intervals were obtained. A second sample of 235 individuals with a diagnosis of ASD was examined to evaluate whether these four OXTR variants were associated with social communication scores on the Autism Diagnostic Interview – Revised (ADI-R). Parameter estimates and associated 95 percent confidence intervals were generated using a linear regression approach. Multiple testing issues were addressed using false discovery adjustments. Results: The rs53576 AG genotype was significantly associated with a lower risk of ASD (OR = 0.707, 95% CI: 0.512-0.975). A single genotype (AG) provided by the rs2254298 marker was found to be significantly associated with higher social communication scores (Parameter estimate = 1.833, SE = 0.762, p = 0.0171). This association was also seen in a Caucasian only and mothers as the respondent samples. No association was significant following false discovery rate adjustments. Conclusion: The findings from these studies provide limited support for the role of OXTR SNPs in ASD, especially in social communication skills. The clinical significance of these associations remains unknown, however, it is likely that these associations do not play a role in the severity of symptoms associated with ASD. Rather, they may be important in the appearance of social deficits due to the rs2254298 markers association with enlarged amygdalas.

Flexible strategies for association analysis with genomic phenotypes

Thesis (Ph.D.)--University of Washington, 2016-08

Genetic Polymorphisms in Replication-related Genes and Risk of Breast Cancer Among European and East Asian Women

Background: The role of common, low to intermediate risk alleles in breast cancer need to be examined due to their relatively high prevalence. Among many cellular pathways, replication has a pivotal role in cell division and frequently targeted during carcinogenesis. Replication is governed by a host of genes involved in a number of different pathways. This study investigates the effects of replication-gene variants in relation to breast cancer and how this relationship is affected by ethnicity, menopausal status and breast tumour subtype. Methods: Data from a case-control study with 997 incident breast cancer cases and 1,050 age frequency matched controls in Vancouver, British Columbia and Kingston, Ontario were used. Unconditional logistic regression was used to calculate odds ratios between 45 replication gene variants and breast cancer risk, assuming an additive genetic model adjusted for age and centre, presented for Europeans and East Asians separately. Polytomous logistic regression was used to assess odds ratios between each SNP and four breast cancer subtypes defined by hormone receptor status among Europeans. All analyses were stratified by menopausal status. The Benjamini–Hochberg false discovery rate (FDR) was used to address multiple comparisons. Results: Among Europeans, the SNPs in FGFR2, TOX3 and 11q13 loci were associated with breast cancer after controlling for multiple comparisons. Test of heterogeneity showed the SNPs rs1045185, rs4973768, rs672888, rs1219648, rs2420946 among Europeans and rs889312 among East Asians conferred differential risk across the tumour subtypes. Conclusions: Specific SNPs in replication genes were associated with breast cancer, and the risk level differed by tumour subtype defined by ER/PR/Her2 status and ethnicity.