Abstract
Spices and herbs are key dietary ingredients used in cuisines across the world. They have been reported to be of medicinal value for a wide variety of diseases through a large body of biomedical investigations. Bioactive phytochemicals in these plant products form the basis of their therapeutic potential as well as adverse effects. A systematic compilation of empirical data involving these aspects of culinary spices and herbs could help unravel molecular mechanisms underlying their effects on health.
SpiceRx provides a platform for exploring the health impact of spices and herbs used in food preparations through a structured database of tripartite relationships with their phytochemicals and disease associations. Starting with an extensive dictionary of culinary spices and herbs, their disease associations were text mined from MEDLINE, the largest database of biomedical abstracts, assisted with manual curation. This information was further combined with spice-phytochemical and phytochemical-disease associations. SpiceRx is an integrated repertoire of evidence-based knowledge pertaining to the health impacts of culinary spices and herbs, and facilitates their disease-specific culinary recommendations as well as exploration of molecular mechanisms underlying their health effects.
Availability and Implementation SpiceRx is available at http://cosylab.iiitd.edu.in/spicerx and supports all modern browsers. SpiceRx is implemented with Python web development framework Django and relational database PostgreSQL; the front-end was built using HTML, CSS, JavaScript, AJAX, jQuery, JSME Molecular Editor, Bootstrap, Jmol, DataTables and Google Charts.
Introduction
Spices and herbs have a unique place in culinary preparations across the world cuisines (Singh and Bagler, unpublished; CulinaryDB, http://cosylab.iiitd.edu.in/culinarydb). While they have been appreciated as flavoring agents and suggested to be of value as antimicrobial agents, a coherent picture of their health effects is hitherto unavailable (Billing and Sherman, 1998). Reductionist investigations have primarily focused on benevolent and adverse effects of individual spices or their phytochemicals, presenting a potpourri of facts with no coherent picture (Srinivasan, 2005b; Yashin et al., 2017; Srinivasan, 2005a). While such studies have added to the fragments of evidence at the levels of culinary spices/herbs and their molecular constituents, a holistic understanding of their health impacts remains vague and unstructured.
Integration of scientific evidence available from an exponentially growing literature reporting health consequences of culinary spices and herbs will enable drawing inferences for their informed culinary use as well as for generating hypotheses to discover underlying molecular mechanisms (Rakhi et al., unpublished). SpiceRx bridges information associating culinary spices/herbs, their phytochemicals and diseases with the help of evidence compiled from research articles and external resources to provide a platform for open-ended explorations of their tripartite relationships (Figure 1).
SpiceRx is an integrated repertoire of evidence-based knowledge pertaining to the health impacts of culinary spices/herbs and their phytochemicals. It seamlessly integrates scientific evidence from biomedical literature and external resources using a text mining protocol.
Database Overview
SpiceRx provides a platform for exploring the health impact of spices and herbs used in food preparations through a structured database of tripartite relationships with their phytochemicals and disease associations (Figure 1). Starting with an extensive dictionary of 188 culinary spices and herbs, their disease associations were text mined from 28 million MEDLINE abstracts. These data were further combined with evidence of spice-phytochemical and phytochemical-disease associations. SpiceRx presents a compilation of 11750 MEDLINE abstracts containing 8957 disease associations (8172 positive and 783 negative) for 152 spices linked to 848 unique disease-specific Medical Subject Headings (MeSH) IDs (Lipscomb, 2000). The hierarchical organization of MeSH was used as a basis for ontological classification of disease terms. Spice names were tagged using a dictionary matching method whereas diseases were tagged using a machine learning based model, TaggerOne (Leaman and Lu, 2016). A convolutional neural network based relation extraction model (Nguyen and Grishman, 2015; Kumar Sahu et al., 2016) was trained on 6712 manually annotated sentences to extract and classify positive, negative and neutral associations.
Information of 866 phytochemicals was obtained for 142 of the spices using KNApSAcK (Afendi et al., 2012) and PhenolExplorer (Rothwell et al., 2013) that comprised of 570 bioactive compounds with a total of 2042 spice-phytochemical associations. These compounds were further linked to diseases with the help of Comparative Toxicogenomics Database (CTD) (Davis et al., 2017), a public database of literature-curated and inferred chemical-disease associations. The resource unearths literature-supported spice-disease associations for which the spice phytochemical(s) have been independently reported for therapeutic effects. Thus, through interlinked triangular relationships between culinary herbs and spices, their phytochemicals, and diseases, SpiceRx facilitates seamless exploration of evidence-based knowledge for their disease-specific culinary recommendations as well as enquiry into underlying molecular mechanisms.
Architecture and Web Interface
SpiceRx has been designed to facilitate explorations starting from either a spice/herb, a disease, or a phytochemical query to find its association with the remaining two elements.
Spice Search
Culinary spices/herbs can be searched using their common name, scientific name or NCBI taxonomy ID to obtain their disease associations and constituent phytochemicals. Searching for a spice yields paginated list of its disease associations (ranked in descending order of number of publications) and that of its phytochemicals. Research articles reporting the association are listed with link-outs to PubMed. For 20% associations, therapeutic spice phytochemicals involved in a spice-disease association discovered through triangular causal linking are listed. These provide a ground for their pharmaceutical and nutraceutical applications in addition to informed culinary use. Separately, a list of all spice-linked phytochemicals is provided. Phytochemicals could be further explored for their detailed physicochemical features, drug-likeness, and links to associated spices and diseases.
Disease Search
To identify culinary spices and herbs that are reported with either therapeutic or adverse effect against a specific disease, SpiceRx provides disease search integrated with the hierarchical organization of MeSH disease terms. One may search by Disease Name, Disease Category, Disease Sub Category or MeSH ID. The Disease Name could be used to search by common names, such as ‘Diabetes‘, ‘Weight Gain‘, or ‘Obesity‘. The ‘MeSH Disease Category’ represents a broad class of diseases such as ‘Cardiovascular Diseases‘, ‘Endocrine System Diseases’ or ‘Neoplasms‘, whereas ‘Disease Sub Category’ represents more refined disease terms such as ‘Heart Diseases‘, ‘Metabolic Diseases‘, ‘Liver Diseases’ etc. A null search performed with no specific disease term is designed to present an exhaustive list of diseases and their spice associations. Details of spices reported with positive or negative effects for the queried disease term along with link-outs to PubMed articles are presented, in addition to any specific phytochemical(s) involved in the therapeutic association, whenever available.
Phytochemical Search
Spice phytochemicals form the basis for molecular mechanisms involved in therapeutic effects against specific diseases. Apart from querying SpiceRx for a spice or a disease, one may also explore the resource to search compounds by their structure, common name, IUPAC name, PubChem ID, molecular weight, Hydrogen bond donors/acceptors or molecular hydrophobicity (AlogP). Each phytochemical could further be explored for its chemical profile (MESH ID, PubChem ID, Common name, IUPAC name, Molecular Formula, Canonical and Isomeric SMILES; Phytochemical and ADMET properties), disease associations as well as a list of spices in which it is reported to be found in. Apart from 2D and 3D visualizations and download options (Mol, 2D Image, and SDF), lookup for structurally similar spice compounds within the database as well as those commercially available from external sources is also provided. The ‘null search’ (with all fields empty) is designed to yield a list of all spice phytochemicals.
Please see Supplementary Information for details of materials and methods, tech stack, database statistics, web interface search features, use cases and data download options.
Conclusions and Discussion
SpiceRx integrates scientific evidence from biomedical literature and external resources to seamlessly collate tripartite associations between culinary spices/herbs, diseases, and phytochemicals. By blending scattered and disorganized evidence, it provides a platform for investigation of spices for their health effects, chemical mechanisms behind their action, and paves way for developing nutraceuticals and drugs.
The effect of spice/herb on a disease may sensitively depends on its quantity. Due to unavailability of uniform data, SpiceRx does not include this information. Further, disease associations can vary according various factors such as the age, pre-existing conditions and gender among others, which are not represented at present. There is scope to enhance SpiceRx to include such additional features.
Funding
No funds were received for this work.
Conflict of Interest
none declared.
Acknowledgements
G.B. thanks the Indraprastha Institute of Information Technology (IIIT-Delhi) for providing computational facilities and support. R.N.K. thanks the Ministry of Human Resource Development, Government of India and Indian Institute of Technology Jodhpur for the senior research fellowship. R.T. (Research Associate), N.G. and J.M. (Research Interns) are affiliated to Dr. Bagler’s lab at the Center for Computational Biology, and are thankful to IIIT-Delhi for the support. G.B. thanks Smita Sudheer, Resmi Vava, Vinay Randhawa and Shreyas Mangave for their feedback to improve the SpiceRx interface.
References
8. References for Supplementary Information
