OBJECTIVE: An increasing challenge in population health research is efficiently utilising the wealth of data available from multiple sources to investigate disease mechanisms and identify potential intervention targets. The use of biomedical data integration platforms can facilitate evidence triangulation from these different sources, improving confidence in causal relationships of interest. In this work, we aimed to integrate Mendelian randomization (MR) and literature-mined evidence from the EpiGraphDB biomedical knowledge graph to build a comprehensive overview of risk factors for developing breast cancer. METHODS: We utilised MR-EvE ("Everything-vs-Everything") data to identify candidate risk factors for breast cancer and generate hypotheses for potential mediators of their effect. We also integrated this data with literature-mined relationships, which were extracted by overlapping literature spaces of risk factors and breast cancer. The literature-based discovery (LBD) results were followed up by validation with two-step MR to triangulate the findings from two data sources. RESULTS: We identified 129 novel and established lifestyle risk factors and molecular traits with evidence of an effect on breast cancer, and made the MR results available in an R/Shiny app (https://mvab.shinyapps.io/MR_heatmaps/). We developed an LBD approach for identifying potential mechanistic intermediates of identified risk factors. We present the results of MR and literature evidence integration for two case studies (childhood body size and HDL-cholesterol), demonstrating their complementary functionalities. CONCLUSION: We demonstrate that MR-EvE data offers an efficient hypothesis-generating approach for identifying disease risk factors. Moreover, we show that integrating MR evidence with literature-mined data may be used to identify causal intermediates and uncover the mechanisms behind the disease.

MOTIVATION: Integrating information from data sources representing different study designs has the potential to strengthen evidence in population health research. However, this concept of evidence "triangulation" presents a number of challenges for systematically identifying and integrating relevant information. These include the harmonization of heterogenous evidence with common semantic concepts and properties, as well as the priortization of the retrieved evidence for triangulation with the question of interest. RESULTS: We present Annotated Semantic Queries (ASQ), a natural language query interface to the integrated biomedical entities and epidemiological evidence in EpiGraphDB, which enables users to extract "claims" from a piece of unstructured text, and then investigate the evidence that could either support, contradict the claims, or offer additional information to the query. This approach has the potential to support the rapid review of preprints, grant applications, conference abstracts, and articles submitted for peer review. ASQ implements strategies to harmonize biomedical entities in different taxonomies and evidence from different sources, to facilitate evidence triangulation and interpretation. AVAILABILITY AND IMPLEMENTATION: ASQ is openly available at https://asq.epigraphdb.org and its source code is available at https://github.com/mrcieu/epigraphdb-asq under GPL-3.0 license.

MOTIVATION: Adverse reactions from drug combinations are increasingly common, making their accurate prediction a crucial challenge in modern medicine. Laboratory-based identification of these reactions is insufficient due to the combinatorial nature of the problem. While many computational approaches have been proposed, tensor factorization (TF) models have shown mixed results, necessitating a thorough investigation of their capabilities when properly optimized. RESULTS: We demonstrate that TF models can achieve state-of-the-art performance on polypharmacy side effect prediction, with our best model (SimplE) achieving median scores of 0.978 area under receiver-operating characteristic curve, 0.971 area under precision-recall curve, and 1.000 AP@50 across 963 side effects. Notably, this model reaches 98.3% of its maximum performance after just two epochs of training (approximately 4 min), making it substantially faster than existing approaches while maintaining comparable accuracy. We also find that incorporating monopharmacy data as self-looping edges in the graph performs marginally better than using it to initialize embeddings. AVAILABILITY AND IMPLEMENTATION: All code used in the experiments is available in our GitHub repository (https://doi.org/10.5281/zenodo.10684402). The implementation was carried out using Python 3.8.12 with PyTorch 1.7.1, accelerated with CUDA 11.4 on NVIDIA GeForce RTX 2080 Ti GPUs.

MOTIVATION: Human traits are typically represented in both the biomedical literature and large population studies as descriptive text strings. Whilst a number of ontologies exist, none of these perfectly represent the entire human phenome and exposome. Mapping trait names across large datasets is therefore time-consuming and challenging. Recent developments in language modelling have created new methods for semantic representation of words and phrases, and these methods offer new opportunities to map human trait names in the form of words and short phrases, both to ontologies and to each other. Here, we present a comparison between a range of established and more recent language modelling approaches for the task of mapping trait names from UK Biobank to the Experimental Factor Ontology (EFO), and also explore how they compare to each other in direct trait-to-trait mapping. RESULTS: In our analyses of 1191 traits from UK Biobank with manual EFO mappings, the BioSentVec model performed best at predicting these, matching 40.3% of the manual mappings correctly. The BlueBERT-EFO model (finetuned on EFO) performed nearly as well (38.8% of traits matching the manual mapping). In contrast, Levenshtein edit distance only mapped 22% of traits correctly. Pairwise mapping of traits to each other demonstrated that many of the models can accurately group similar traits based on their semantic similarity. AVAILABILITY AND IMPLEMENTATION: Our code is available at https://github.com/MRCIEU/vectology.

Bioinformatics (2020) doi: 10.1093/bioinformatics/btaa961 Upon the original publication of this article, there was an error in the source code syntax under sub-section “2.2 Integration of epidemiological evidence” in the “Materials and methods” section. The source code syntax should read: “(e.g. (Gwas {trait: ‘Body mass index’})-[MR {beta, se, pval}]->(Gwas {trait: ‘Coronary heart disease’}))” instead of “ (e.g. [Gwas (trait: ‘Body mass index’)]-[MR {beta, se, pval}]->(Gwas {trait: ‘Coronary heart disease’})))”. This error has now been corrected. The Publisher apologizes for the error.

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.