Exact statistical tests for the intersection of independent lists of genes

miRNA dysregulation in traumatic brain injury and epilepsy: a systematic review to identify putative biomarkers for post-traumatic epilepsy

Traumatic brain injury (TBI) leads to post-traumatic epilepsy (PTE); hence, both TBI and PTE share various similar molecular mechanisms. MicroRNA (miRNA) is a small noncoding RNA that acts as a gene-silencing molecule. Notably, the dysregulation of miRNAs in various neurological diseases, including TBI and epilepsy, has been reported in several studies. However, studies on commonly dysregulated miRNAs and the regulation of shared pathways in both TBI and epilepsy that can identify potential biomarkers of PTE are still lacking. This systematic review covers the peer-review publications of TBI and database studies of epilepsy-dysregulated miRNAs of clinical studies. For TBI, 290 research articles were identified after screening, and 12 provided data for dysregulated miRNAs in humans. The compiled data suggest that 85 and 222 miRNAs are consecutively dysregulated in TBI and epilepsy. In both, 10 miRNAs were found to be commonly dysregulated, implying that they are potentially dysregulated miRNAs for PTE. Furthermore, the targets and involvement of each putative miRNA in different pathways were identified and evaluated. Additionally, clusters of predicted miRNAs were analyzed. Each miRNA's regulatory role was linked with apoptosis, inflammation, and cell cycle regulation pathways. Hence, these findings provide insight for future diagnostic biomarkers.

Querying multiple sets of P-values through composed hypothesis testing

MOTIVATION

Combining the results of different experiments to exhibit complex patterns or to improve statistical power is a typical aim of data integration. The starting point of the statistical analysis often comes as a set of P-values resulting from previous analyses, that need to be combined flexibly to explore complex hypotheses, while guaranteeing a low proportion of false discoveries.

RESULTS

We introduce the generic concept of composed hypothesis, which corresponds to an arbitrary complex combination of simple hypotheses. We rephrase the problem of testing a composed hypothesis as a classification task and show that finding items for which the composed null hypothesis is rejected boils down to fitting a mixture model and classifying the items according to their posterior probabilities. We show that inference can be efficiently performed and provide a thorough classification rule to control for type I error. The performance and the usefulness of the approach are illustrated in simulations and on two different applications. The method is scalable, does not require any parameter tuning, and provided valuable biological insight on the considered application cases.

AVAILABILITY AND IMPLEMENTATION

The QCH methodology is available in the qch package hosted on CRAN. Additionally, R codes to reproduce the Einkorn example are available on the personal webpage of the first author: https://www6.inrae.fr/mia-paris/Equipes/Membres/Tristan-Mary-Huard.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Substantial Increase in Compliance with Saturated Fatty Acid Intake Recommendations after One Year Following the American Heart Association Diet

The American Heart Association (AHA) dietary guidelines recommend 30–35% of energy intake (%E) be from total fat, <7%E from saturated fatty acids (SFA), and <1%E from trans fatty acid (TFA). This study evaluates the effect of AHA dietary counselling on fat intake. Between 2009 and 2014, 119 obese adults with metabolic syndrome (MetS), (71% women, average 52.5 years of age, and 34.9 kg/m² of body mass index), received individual and group counselling on the AHA diet, over a one-year study period. Each participant attended 2 individual sessions (months 1 and 12) and 12 group sessions, at one-month intervals. At baseline and one-year, we collected three random 24-h diet recalls (two weekdays and one weekend day). Fat intake patterns over time were analyzed using paired-t test and linear mixed-effect models. There was significant variation on SFA and TFA intake per meal, being highest at dinner, in restaurants, and on weekends. Over the one-year study period, daily intake of total fat, SFA, and TFA decreased by 27%, 37% and 41%, respectively (p-value < 0.01, each). Correspondingly, the percentage of participants complying with AHA’s recommendations, increased from 25.2% to 40.2% for total fat (p-value = 0.02); from 2.5% to 20.7% for SFA (p-value < 0.01); and from 45.4% to 62% for TFA (p-value = 0.02). Additionally, SFA intake for all meal types at home decreased significantly (p-value < 0.05, each). AHA dietary counselling significantly increased the compliance with AHA dietary guidelines, with an eightfold increase in compliance in SFA intake. Nonetheless, ~80% of our participants still exceeded the recommended SFA intake. Substantial efforts are needed to encourage low-SFA and low-TFA food preparation at home, with strong public health policies to decrease SFA and TFA in restaurants and prepared foods.

Comprehensive analysis of normal adjacent to tumor transcriptomes

Histologically normal tissue adjacent to the tumor (NAT) is commonly used as a control in cancer studies. However, little is known about the transcriptomic profile of NAT, how it is influenced by the tumor, and how the profile compares with non-tumor-bearing tissues. Here, we integrate data from the Genotype-Tissue Expression project and The Cancer Genome Atlas to comprehensively analyze the transcriptomes of healthy, NAT, and tumor tissues in 6506 samples across eight tissues and corresponding tumor types. Our analysis shows that NAT presents a unique intermediate state between healthy and tumor. Differential gene expression and protein–protein interaction analyses reveal altered pathways shared among NATs across tissue types. We characterize a set of 18 genes that are specifically activated in NATs. By applying pathway and tissue composition analyses, we suggest a pan-cancer mechanism of pro-inflammatory signals from the tumor stimulates an inflammatory response in the adjacent endothelium.

Normal tissue adjacent to the tumour (NAT) is often used as a control in cancer studies. Here, the authors analyse across cancer types the transcriptomes of healthy, NAT, and tumour tissues, and find that NAT presents a unique state, potentially due to inflammatory response of the NAT to the tumour tissue.

Analogous mechanism regulating formation of neocortical basal radial glia and cerebellar Bergmann glia

Neocortical basal radial glia (bRG) and cerebellar Bergmann glia (BG) are basal progenitors derived from ventricular apical radial glia (aRG) that selectively lose their apical processes. bRG and BG have been implicated in the expansion and folding of the cerebrum and cerebellum, respectively. Here, we analyzed the molecular characteristics and development of bRG and BG. Transcriptomic comparison revealed striking similarity of the molecular features of bRG and BG. We found that heightened ERK signaling activity in aRG is tightly linked to the temporal formation and the relative abundance of bRG in human and mouse cortices. Forced activation of an FGF-ERK-ETV axis that is crucial to BG induction specifically induced bRG with canonical human bRG features in mice. Therefore, our data point to a common mechanism of bRG and BG generation, bearing implications to the role for these basal progenitors in the evolution of cortical folding of the cerebrum and cerebellum.

DOI:http://dx.doi.org/10.7554/eLife.23253.001

The outer layer of the brain of a mammal, called the cortex, helps support mental abilities such as memory, attention and thought. In rodents, the cortex is smooth whereas in primates it is organized into folds. These folds increase the surface area of the brain and thus the number of neurons it can contain, which may in turn increase its processing power. Folding occurs as the brain develops in the womb. Specialized cells called basal or outer radial glia, which are more abundant in humans than in rodents, are believed to trigger the folding process.

Another area of the brain, called the cerebellum, is intricately folded in both rodents and humans. As the brain develops, cells within the cerebellum called Bergmann glia cause the tissue to fold. Bergmann glia and basal radial glia share a number of similarities, but it was not known whether the same molecular pathway might regulate both types of cell.

Now, Heng et al. show that Bergmann glia in the cerebellums of mice and basal radial glia in human cortex contain similar sets of active genes. Moreover, the molecular pathway that gives rise to Bergmann glia in mice is also active in the cortex of both mice and humans. However, it is much more active in humans, leading Heng et al. to speculate that high levels of activity in this pathway might give rise to basal radial glia. Consistent with this prediction, artificially activating the pathway at high levels in mouse cortex triggered the formation of basal radial glia in mice too. These results thus suggest that a common mechanism generates both types of glial cells involved in brain folding.

The work of Heng et al. lays the foundations for further studies into how these cells fold the brain and thus how they contribute to more complex mental abilities. Remaining questions to address are whether other species with Bergmann glia also have folded cerebellums, and whether incorrect development of basal radial glia in humans leads to disorders in which the cortex folds abnormally.

DOI:http://dx.doi.org/10.7554/eLife.23253.002

Racial Differences in Misclassification of Healthy Eating Based on Food Frequency Questionnaire and 24-Hour Dietary Recalls

OBJECTIVES

To examine the agreement in nutrient intake and alternate healthy eating indices (AHEI) between a self-administered Food Frequency Questionnaire (FFQ) and 24-hour recall (24HR) measurements of diet by race, among urban older women.

DESIGN

Cross-sectional observational study.

SETTING

Urban neighborhoods in Washington, DC, USA.

PARTICIPANTS

Community-dwelling White and Black women aged 65 and older.

MEASUREMENTS

In 2014 and 2015, 49 White and 44 Black older women were queried on diet using both FFQ and 24-hour recalls. The correlation coefficients of 55 nutrient intake measures and agreements on healthy eating classification between the two instruments were compared overall and by race.

RESULTS

The mean correlation coefficient (rho) was 0.46 for Whites and 0.23 for Blacks. For 47 measures, rho was lower for Blacks. Whites had a strong correlation of ≥0.5 for 28 items, while Blacks had strong correlations for only 3 items. Based on FFQ, the mean (SD) of AHEI were 54.0 (10.3) for Whites and 45.9 (8.8) for Blacks (p<0.001). Based on 24HR, the mean (SD) were 43.9 (10.8) for Whites and 33.2 (9.6) for Blacks (p<0.001). Using 32 as the cutoff (40% of maximum AHEI score), 50% of Blacks and 14% of Whites were classified as eating unhealthy based on the 24HR, versus 2.6% and 0% based on the FFQ.

CONCLUSION

The FFQ has limited ability to accurately assess nutrient intake among older Black women, and tends to underestimate racial differences in healthy eating. The FFQ should be further improved for use in racial disparities research of healthy eating in older age, using a larger sample of older women with racial and geographic diversities.

Strategies for Integrated Analysis of Genetic, Epigenetic, and Gene Expression Variation in Cancer: Addressing the Challenges

The development and progression of cancer, a collection of diseases with complex genetic architectures, is facilitated by the interplay of multiple etiological factors. This complexity challenges the traditional single-platform study design and calls for an integrated approach to data analysis. However, integration of heterogeneous measurements of biological variation is a non-trivial exercise due to the diversity of the human genome and the variety of output data formats and genome coverage obtained from the commonly used molecular platforms. This review article will provide an introduction to integration strategies used for analyzing genetic risk factors for cancer. We critically examine the ability of these strategies to handle the complexity of the human genome and also accommodate information about the biological and functional interactions between the elements that have been measured-making the assessment of disease risk against a composite genomic factor possible. The focus of this review is to provide an overview and introduction to the main strategies and to discuss where there is a need for further development.

Efficient Test and Visualization of Multi-Set Intersections

Identification of sets of objects with shared features is a common operation in all disciplines. Analysis of intersections among multiple sets is fundamental for in-depth understanding of their complex relationships. However, so far no method has been developed to assess statistical significance of intersections among three or more sets. Moreover, the state-of-the-art approaches for visualization of multi-set intersections are not scalable. Here, we first developed a theoretical framework for computing the statistical distributions of multi-set intersections based upon combinatorial theory, and then accordingly designed a procedure to efficiently calculate the exact probabilities of multi-set intersections. We further developed multiple efficient and scalable techniques to visualize multi-set intersections and the corresponding intersection statistics. We implemented both the theoretical framework and the visualization techniques in a unified R software package, SuperExactTest. We demonstrated the utility of SuperExactTest through an intensive simulation study and a comprehensive analysis of seven independently curated cancer gene sets as well as six disease or trait associated gene sets identified by genome-wide association studies. We expect SuperExactTest developed by this study will have a broad range of applications in scientific data analysis in many disciplines.

The transcription factor Cux1 in cerebellar granule cell development and medulloblastoma pathogenesis

Cux1, also known as Cutl1, CDP or Cut is a homeodomain transcription factor implicated in the regulation of normal and oncogenic development in diverse peripheral tissues and organs. We studied the expression and functional role of Cux1 in cerebellar granule cells and medulloblastoma. Cux1 is robustly expressed in proliferating granule cell precursors and in postmitotic, migrating granule cells. Expression is lost as postmigratory granule cells mature. Moreover, Cux1 is also strongly expressed in a well-established mouse model of medulloblastoma. In contrast, expression of CUX1 in human medulloblastoma tissue samples is lower than in normal fetal cerebellum. In these tumors, CUX1 expression tightly correlates with a set of genes which, when mapped on a global protein-protein interaction dataset, yields a tight network that constitutes a cell cycle control signature and may be related to p53 and the DNA damage response pathway. Antisense-mediated reduction of CUX1 levels in two human medulloblastoma cell lines led to a decrease in proliferation and altered motility. The developmental expression of Cux1 in the cerebellum and its action in cell lines support a role in granule cell and medulloblastoma proliferation. Its expression in human medulloblastoma shifts that perspective, suggesting that CUX1 is part of a network involved in cell cycle control and maintenance of DNA integrity. The constituents of this network may be rational targets to therapeutically approach medulloblastomas.

BRAIN NETWORKS. Correlated gene expression supports synchronous activity in brain networks

During rest, brain activity is synchronized between different regions widely distributed throughout the brain, forming functional networks. However, the molecular mechanisms supporting functional connectivity remain undefined. We show that functional brain networks defined with resting-state functional magnetic resonance imaging can be recapitulated by using measures of correlated gene expression in a post mortem brain tissue data set. The set of 136 genes we identify is significantly enriched for ion channels. Polymorphisms in this set of genes significantly affect resting-state functional connectivity in a large sample of healthy adolescents. Expression levels of these genes are also significantly associated with axonal connectivity in the mouse. The results provide convergent, multimodal evidence that resting-state functional networks correlate with the orchestrated activity of dozens of genes linked to ion channel activity and synaptic function.

Finding genetic overlaps among diseases based on ranked gene lists

To understand disease relationships in terms of their genetic mechanisms, it is important to study the common genetic basis among different diseases. Although discoveries on pleiotropic genes related to multiple diseases abound, methods flexibly applicable to various types of datasets generated from different studies or experiments are needed to gain big pictures on the genetic relationships among a large number of diseases. We develop a set of genetic similarity measures to gauge the genetic overlap between diseases, as well as several estimators of the number of overlapping disease genes between diseases. These methods are based on ranked gene lists so that they could be flexibly applied to different types of data. We first investigate the performance of the genetic similarity measure for evaluating the similarity between human diseases in simulation studies. Then we apply the method to diseases in the OMIM database. We show that our proposed genetic measure achieves superior performance in explaining phenotype similarities between diseases compared to simpler methods. Furthermore, we identified common genes underlying the genetic overlap between disease pairs. With an example of five vision-related diseases, we demonstrate how our methods can provide insights into the relationships among diseases based on their shared genetic mechanisms.

Deciding whether follow-up studies have replicated findings in a preliminary large-scale omics study

We propose a formal method to declare that findings from a primary study have been replicated in a follow-up study. Our proposal is appropriate for primary studies that involve large-scale searches for rare true positives (i.e., needles in a haystack). Our proposal assigns an r value to each finding; this is the lowest false discovery rate at which the finding can be called replicated. Examples are given and software is available.

Metabolomics reveals signature of mitochondrial dysfunction in diabetic kidney disease

Diabetic kidney disease is the leading cause of ESRD, but few biomarkers of diabetic kidney disease are available. This study used gas chromatography-mass spectrometry to quantify 94 urine metabolites in screening and validation cohorts of patients with diabetes mellitus (DM) and CKD(DM+CKD), in patients with DM without CKD (DM-CKD), and in healthy controls. Compared with levels in healthy controls, 13 metabolites were significantly reduced in the DM+CKD cohorts (P≤0.001), and 12 of the 13 remained significant when compared with the DM-CKD cohort. Many of the differentially expressed metabolites were water-soluble organic anions. Notably, organic anion transporter-1 (OAT1) knockout mice expressed a similar pattern of reduced levels of urinary organic acids, and human kidney tissue from patients with diabetic nephropathy demonstrated lower gene expression of OAT1 and OAT3. Analysis of bioinformatics data indicated that 12 of the 13 differentially expressed metabolites are linked to mitochondrial metabolism and suggested global suppression of mitochondrial activity in diabetic kidney disease. Supporting this analysis, human diabetic kidney sections expressed less mitochondrial protein, urine exosomes from patients with diabetes and CKD had less mitochondrial DNA, and kidney tissues from patients with diabetic kidney disease had lower gene expression of PGC1α (a master regulator of mitochondrial biogenesis). We conclude that urine metabolomics is a reliable source for biomarkers of diabetic complications, and our data suggest that renal organic ion transport and mitochondrial function are dysregulated in diabetic kidney disease.

Exact statistical tests for the intersection of independent lists of genes

Public data repositories have enabled researchers to compare results across multiple genomic studies in order to replicate findings. A common approach is to first rank genes according to an hypothesis of interest within each study. Then, lists of the top-ranked genes within each study are compared across studies. Genes recaptured as highly ranked (usually above some threshold) in multiple studies are considered to be significant. However, this comparison strategy often remains informal, in that Type I error and false discovery rate are usually uncontrolled. In this paper, we formalize an inferential strategy for this kind of list-intersection discovery test. We show how to compute a p-value associated with a `recaptured' set of genes, using a closed-form Poisson approximation to the distribution of the size of the recaptured set. The distribution of the test statistic depends on the rank threshold and the number of studies within which a gene must be recaptured. We use a Poisson approximation to investigate operating characteristics of the test. We give practical guidance on how to design a bioinformatic list-intersection study with prespecified control of Type I error (at the set level) and false discovery rate (at the gene level). We show how choice of test parameters will affect the expected proportion of significant genes identified. We present a strategy for identifying optimal choice of parameters, depending on the particular alternative hypothesis which might hold. We illustrate our methods using prostate cancer gene-expression datasets from the curated Oncomine database.