Search and visualization of gene-drug-disease interactions for pharmacogenomics and precision medicine research using GeneDive

BACKGROUND

Understanding the relationships between genes, drugs, and disease states is at the core of pharmacogenomics. Two leading approaches for identifying these relationships in medical literature are: human expert led manual curation efforts, and modern data mining based automated approaches. The former generates small amounts of high-quality data, and the latter offers large volumes of mixed quality data. The algorithmically extracted relationships are often accompanied by supporting evidence, such as, confidence scores, source articles, and surrounding contexts (excerpts) from the articles, that can be used as data quality indicators. Tools that can leverage these quality indicators to help the user gain access to larger and high-quality data are needed.

APPROACH

We introduce GeneDive, a web application for pharmacogenomics researchers and precision medicine practitioners that makes gene, disease, and drug interactions data easily accessible and usable. GeneDive is designed to meet three key objectives: (1) provide functionality to manage information-overload problem and facilitate easy assimilation of supporting evidence, (2) support longitudinal and exploratory research investigations, and (3) offer integration of user-provided interactions data without requiring data sharing.

RESULTS

GeneDive offers multiple search modalities, visualizations, and other features that guide the user efficiently to the information of their interest. To facilitate exploratory research, GeneDive makes the supporting evidence and context for each interaction readily available and allows the data quality threshold to be controlled by the user as per their risk tolerance level. The interactive search-visualization loop enables relationship discoveries between diseases, genes, and drugs that might not be explicitly described in literature but are emergent from the source medical corpus and deductive reasoning. The ability to utilize user's data either in combination with the GeneDive native datasets or in isolation promotes richer data-driven exploration and discovery. These functionalities along with GeneDive's applicability for precision medicine, bringing the knowledge contained in biomedical literature to bear on particular clinical situations and improving patient care, are illustrated through detailed use cases.

CONCLUSION

GeneDive is a comprehensive, broad-use biological interactions browser. The GeneDive application and information about its underlying system architecture are available at http://www.genedive.net. GeneDive Docker image is also available for download at this URL, allowing users to (1) import their own interaction data securely and privately; and (2) generate and test hypotheses across their own and other datasets.

Comparing tuberculosis gene signatures in malnourished individuals using the TBSignatureProfiler

Background

Gene expression signatures have been used as biomarkers of tuberculosis (TB) risk and outcomes. Platforms are needed to simplify access to these signatures and determine their validity in the setting of comorbidities. We developed a computational profiling platform of TB signature gene sets and characterized the diagnostic ability of existing signature gene sets to differentiate active TB from LTBI in the setting of malnutrition.

Methods

We curated 45 existing TB-related signature gene sets and developed our TBSignatureProfiler software toolkit that estimates gene set activity using multiple enrichment methods and allows visualization of single- and multi-pathway results. The TBSignatureProfiler software is available through Bioconductor and on GitHub. For evaluation in malnutrition, we used whole blood gene expression profiling from 23 severely malnourished Indian individuals with TB and 15 severely malnourished household contacts with latent TB infection (LTBI). Severe malnutrition was defined as body mass index (BMI) < 16 kg/m2 in adults and based on weight-for-height Z scores in children < 18 years. Gene expression was measured using RNA-sequencing.

Results

The comparison and visualization functions from the TBSignatureProfiler showed that TB gene sets performed well in malnourished individuals; 40 gene sets had statistically significant discriminative power for differentiating TB from LTBI, with area under the curve ranging from 0.662–0.989. Three gene sets were not significantly predictive.

Conclusion

Our TBSignatureProfiler is a highly effective and user-friendly platform for applying and comparing published TB signature gene sets. Using this platform, we found that existing gene sets for TB function effectively in the setting of malnutrition, although differences in gene set applicability exist. RNA-sequencing gene sets should consider comorbidities and potential effects on diagnostic performance.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-020-05598-z.

A primer for generating and using transcriptome data and gene sets

Transcriptomic approaches have provided a growing set of powerful tools with which to study genome-wide patterns of gene expression. Rapidly evolving technologies enable analysis of transcript abundance data from particular tissues and even single cells. This Primer discusses methods that can be used to collect and profile RNAs from specific tissues or cells, process and analyze high-throughput RNA-sequencing data, and define sets of genes that accurately represent a category, such as tissue-enriched or tissue-specific gene expression.

Summary: This Primer gives an overview of the considerations required when planning and conducting transcriptome profiling experiments, from sample isolation to data analysis and display, and discusses best practice in defining gene sets for use in addressing biological questions.

Pathway information extracted from 25 years of pathway figures

Thousands of pathway diagrams are published each year as static figures inaccessible to computational queries and analyses. Using a combination of machine learning, optical character recognition, and manual curation, we identified 64,643 pathway figures published between 1995 and 2019 and extracted 1,112,551 instances of human genes, comprising 13,464 unique NCBI genes, participating in a wide variety of biological processes. This collection represents an order of magnitude more genes than found in the text of the same papers, and thousands of genes missing from other pathway databases, thus presenting new opportunities for discovery and research.

Muscle Gene Sets: a versatile methodological aid to functional genomics in the neuromuscular field

BACKGROUND

The approach of building large collections of gene sets and then systematically testing hypotheses across these collections is a powerful tool in functional genomics, both in the pathway analysis of omics data and to uncover the polygenic effects associated with complex diseases in genome-wide association study. The Molecular Signatures Database includes collections of oncogenic and immunologic signatures enabling researchers to compare transcriptional datasets across hundreds of previous studies and leading to important insights in these fields, but such a resource does not currently exist for neuromuscular research. In previous work, we have shown the utility of gene set approaches to understand muscle cell physiology and pathology.

METHODS

Following a systematic survey of public muscle data, we passed gene expression profiles from 4305 samples through a robust pre-processing and standardized data analysis pipeline. Two hundred eighty-two samples were discarded based on a battery of rigorous global quality controls. From among the remaining studies, 578 comparisons of interest were identified by a combination of text mining and manual curation of the study meta-data. For each comparison, significantly dysregulated genes (FDR adjusted p < 0.05) were identified.

RESULTS

Lists of dysregulated genes were divided between upregulated and downregulated to give 1156 Muscle Gene Sets (MGS). This resource is available for download ( www.sys-myo.com/muscle_gene_sets ) and is accessible through three commonly used functional genomics platforms (GSEA, EnrichR, and WebGestalt). Basic guidance and recommendations are provided for the use of MGS through these platforms. In addition, consensus muscle gene sets were created to capture the overlap between the results of similar studies, and analysis of these highlighted the potential for novel disease-relevant findings.

CONCLUSIONS

The MGS resource can be used to investigate the behaviour of any list of genes across previous comparisons of muscle conditions, to compare previous studies to one another, and to explore the functional relationship of muscle dysregulation to the Gene Ontology. Its major intended use is in enrichment testing for functional genomics analysis.

Genes2GO: A web application for querying gene sets for specific GO terms

Gene ontology annotations have become an essential resource for biological interpretations of experimental findings. The process of gathering basic annotation information in tables that link gene sets with specific gene ontology terms can be cumbersome, in particular if it requires above average computer skills or bioinformatics expertise. We have therefore developed Genes2GO, an intuitive R-based web application. Genes2GO uses the biomaRt package of Bioconductor in order to retrieve custom sets of gene ontology annotations for any list of genes from organisms covered by the Ensembl database. Genes2GO produces a binary matrix file, indicating for each gene the presence or absence of specific annotations for a gene. It should be noted that other GO tools do not offer this user-friendly access to annotations.

Empirical pathway analysis, without permutation

Resampling-based expression pathway analysis techniques have been shown to preserve type I error rates, in contrast to simple gene-list approaches that implicitly assume the independence of genes in ranked lists. However, resampling is intensive in computation time and memory requirements. We describe accurate analytic approximations to permutations of score statistics, including novel approaches for Pearson's correlation, and summed score statistics, that have good performance for even relatively small sample sizes. Our approach preserves the essence of permutation pathway analysis, but with greatly reduced computation. Extensions for inclusion of covariates and censored data are described, and we test the performance of our procedures using simulations based on real datasets. These approaches have been implemented in the new R package safeExpress.