EpiGraphDB: a database and data mining platform for health data science.

MOTIVATION: The wealth of data resources on human phenotypes, risk factors, molecular traits and therapeutic interventions presents new opportunities for population health sciences. These opportunities are paralleled by a growing need for data integration, curation and mining to increase research efficiency, reduce mis-inference and ensure reproducible research. RESULTS: We developed EpiGraphDB (https://epigraphdb.org/), a graph database containing an array of different biomedical and epidemiological relationships and an analytical platform to support their use in human population health data science. In addition, we present three case studies that illustrate the value of this platform. The first uses EpiGraphDB to evaluate potential pleiotropic relationships, addressing mis-inference in systematic causal analysis. In the second case study, we illustrate how protein-protein interaction data offer opportunities to identify new drug targets. The final case study integrates causal inference using Mendelian randomization with relationships mined from the biomedical literature to 'triangulate' evidence from different sources. AVAILABILITY AND IMPLEMENTATION: The EpiGraphDB platform is openly available at https://epigraphdb.org. Code for replicating case study results is available at https://github.com/MRCIEU/epigraphdb as Jupyter notebooks using the API, and https://mrcieu.github.io/epigraphdb-r using the R package. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Model-based clustering of multi-tissue gene expression data.

MOTIVATION: Recently, it has become feasible to generate large-scale, multi-tissue gene expression data, where expression profiles are obtained from multiple tissues or organs sampled from dozens to hundreds of individuals. When traditional clustering methods are applied to this type of data, important information is lost, because they either require all tissues to be analyzed independently, ignoring dependencies and similarities between tissues, or to merge tissues in a single, monolithic dataset, ignoring individual characteristics of tissues. RESULTS: We developed a Bayesian model-based multi-tissue clustering algorithm, revamp, which can incorporate prior information on physiological tissue similarity, and which results in a set of clusters, each consisting of a core set of genes conserved across tissues as well as differential sets of genes specific to one or more subsets of tissues. Using data from seven vascular and metabolic tissues from over 100 individuals in the STockholm Atherosclerosis Gene Expression (STAGE) study, we demonstrate that multi-tissue clusters inferred by revamp are more enriched for tissue-dependent protein-protein interactions compared to alternative approaches. We further demonstrate that revamp results in easily interpretable multi-tissue gene expression associations to key coronary artery disease processes and clinical phenotypes in the STAGE individuals. AVAILABILITY AND IMPLEMENTATION: Revamp is implemented in the Lemon-Tree software, available at https://github.com/eb00/lemon-tree. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Quantitative analysis of somatically acquired and constitutive uniparental disomy in gastrointestinal cancers.

Somatically acquired uniparental disomies (aUPDs) are frequent events in solid tumors and have been associated with cancer-related genes. Studies assessing their functional consequences across several cancer types are therefore necessary. Here, we aimed at integrating aUPD profiles with the mutational status of cancer-related genes in a tumor-type specific manner. Using TCGA datasets for 1,032 gastrointestinal cancers, including colon (COAD), rectum (READ), stomach (STAD), esophageal adenocarcinoma (EAC) and esophageal squamous cell carcinoma (ESCC), we show a non-random distribution of aUPD, suggesting the existence of a cancer-specific landscape of aUPD events. Our analysis indicates that aUPD acts as a "second hit" in Knudson's model in order to achieve biallelic inactivation of tumor suppressor genes. In particular, APC, ARID1A and NOTCH1 were recurrently inactivated by the presence of homozygous mutation as a consequence of aUPD in COAD and READ, STAD and ESCC, respectively. Furthermore, while TP53 showed inactivation caused by aUPD at chromosome arm 17p across all tumor types, copy number losses at this genomic position were also frequent. By experimental and computationally inferring genome ploidy, we demonstrate that an increased number of aUPD events, both affecting the whole chromosome or segments of it, were present in highly aneuploid genomes compared to near-diploid tumors. Finally, the presence of mosaic UPD was detected at a higher frequency in DNA extracted from peripheral blood lymphocytes of patients with colorectal cancer compared to healthy individuals. In summary, our study defines specific profiles of aUPD in gastrointestinal cancers and provides unequivocal evidence of their relevance in cancer.

Changes in the gene expression profile during spontaneous migraine attacks.

Migraine attacks are delimited, allowing investigation of changes during and outside attack. Gene expression fluctuates according to environmental and endogenous events and therefore, we hypothesized that changes in RNA expression during and outside a spontaneous migraine attack exist which are specific to migraine. Twenty-seven migraine patients were assessed during a spontaneous migraine attack, including headache characteristics and treatment effect. Blood samples were taken during attack, two hours after treatment, on a headache-free day and after a cold pressor test. RNA-Sequencing, genotyping, and steroid profiling were performed. RNA-Sequences were analyzed at gene level (differential expression analysis) and at network level, and genomic and transcriptomic data were integrated. We found 29 differentially expressed genes between 'attack' and 'after treatment', after subtracting non-migraine specific genes, that were functioning in fatty acid oxidation, signaling pathways and immune-related pathways. Network analysis revealed mechanisms affected by changes in gene interactions, e.g. 'ion transmembrane transport'. Integration of genomic and transcriptomic data revealed pathways related to sumatriptan treatment, i.e. '5HT1 type receptor mediated signaling pathway'. In conclusion, we uniquely investigated intra-individual changes in gene expression during a migraine attack. We revealed both genes and pathways potentially involved in the pathophysiology of migraine and/or migraine treatment.

Phenome-wide Mendelian randomization mapping the influence of the plasma proteome on complex diseases.

The human proteome is a major source of therapeutic targets. Recent genetic association analyses of the plasma proteome enable systematic evaluation of the causal consequences of variation in plasma protein levels. Here we estimated the effects of 1,002 proteins on 225 phenotypes using two-sample Mendelian randomization (MR) and colocalization. Of 413 associations supported by evidence from MR, 130 (31.5%) were not supported by results of colocalization analyses, suggesting that genetic confounding due to linkage disequilibrium is widespread in naïve phenome-wide association studies of proteins. Combining MR and colocalization evidence in cis-only analyses, we identified 111 putatively causal effects between 65 proteins and 52 disease-related phenotypes ( https://www.epigraphdb.org/pqtl/ ). Evaluation of data from historic drug development programs showed that target-indication pairs with MR and colocalization support were more likely to be approved, evidencing the value of this approach in identifying and prioritizing potential therapeutic targets.

Learning Differential Module Networks Across Multiple Experimental Conditions.

Module network inference is a statistical method to reconstruct gene regulatory networks, which uses probabilistic graphical models to learn modules of coregulated genes and their upstream regulatory programs from genome-wide gene expression and other omics data. Here, we review the basic theory of module network inference, present protocols for common gene regulatory network reconstruction scenarios based on the Lemon-Tree software, and show, using human gene expression data, how the software can also be applied to learn differential module networks across multiple experimental conditions.

Energy balance of triathletes during an ultra-endurance event.

UNLABELLED: The nutritional strategy during an ultra-endurance triathlon (UET) is one of the main concerns of athletes competing in such events. The purpose of this study is to provide a proper characterization of the energy and fluid intake during real competition in male triathletes during a complete UET and to estimate the energy expenditure (EE) and the fluid balance through the race. METHODS: Eleven triathletes performed a UET. All food and drinks ingested during the race were weighed and recorded in order to assess the energy intake (EI) during the race. The EE was estimated from heart rate (HR) recordings during the race, using the individual HR-oxygen uptake (Vo2) regressions developed from three incremental tests on the 50-m swimming pool, cycle ergometer, and running treadmill. Additionally, body mass (BM), total body water (TBW) and intracellular (ICW) and extracellular water (ECW) were assessed before and after the race using a multifrequency bioimpedance device (BIA). RESULTS: Mean competition time and HR was 755 ± 69 min and 137 ± 6 beats/min, respectively. Mean EI was 3643 ± 1219 kcal and the estimated EE was 11,009 ± 664 kcal. Consequently, athletes showed an energy deficit of 7365 ± 1286 kcal (66.9% ± 11.7%). BM decreased significantly after the race and significant losses of TBW were found. Such losses were more related to a reduction of extracellular fluids than intracellular fluids. CONCLUSIONS: Our results confirm the high energy demands of UET races, which are not compensated by nutrient and fluid intake, resulting in a large energy deficit.

Physiological responses in relation to performance during competition in elite synchronized swimmers.

PURPOSE: We aimed to characterize the cardiovascular, lactate and perceived exertion responses in relation to performance during competition in junior and senior elite synchronized swimmers. METHODS: 34 high level senior (21.4 ± 3.6 years) and junior (15.9 ± 1.0) synchronized swimmers were monitored while performing a total of 96 routines during an official national championship in the technical and free solo, duet and team competitive programs. Heart rate was continuously monitored. Peak blood lactate was obtained from serial capillary samples during recovery. Post-exercise rate of perceived exertion was assessed using the Borg CR-10 scale. Total competition scores were obtained from official records. RESULTS: Data collection was complete in 54 cases. Pre-exercise mean heart rate (beats·min(-1)) was 129.1 ± 13.1, and quickly increased during the exercise to attain mean peak values of 191.7 ± 8.7, with interspersed bradycardic events down to 88.8 ± 28.5. Mean peak blood lactate (mmol·L(-1)) was highest in the free solo (8.5 ± 1.8) and free duet (7.6 ± 1.8) and lowest at the free team (6.2 ± 1.9). Mean RPE (0-10+) was higher in juniors (7.8 ± 0.9) than in seniors (7.1 ± 1.4). Multivariate analysis revealed that heart rate before and minimum heart rate during the routine predicted 26% of variability in final total score. CONCLUSIONS: Cardiovascular responses during competition are characterized by intense anticipatory pre-activation and rapidly developing tachycardia up to maximal levels with interspersed periods of marked bradycardia during the exercise bouts performed in apnea. Moderate blood lactate accumulation suggests an adaptive metabolic response as a result of the specific training adaptations attributed to influence of the diving response in synchronized swimmers. Competitive routines are perceived as very to extremely intense, particularly in the free solo and duets. The magnitude of anticipatory heart rate activation and bradycardic response appear to be related to performance variability.