key: cord-0838456-tqv9kbb2 authors: del Toro, Noemi; Shrivastava, Anjali; Ragueneau, Eliot; Meldal, Birgit; Combe, Colin; Barrera, Elisabet; Perfetto, Livia; How, Karyn; Ratan, Prashansa; Shirodkar, Gautam; Lu, Odilia; Mészáros, Bálint; Watkins, Xavier; Pundir, Sangya; Licata, Luana; Iannuccelli, Marta; Pellegrini, Matteo; Martin, Maria Jesus; Panni, Simona; Duesbury, Margaret; Vallet, Sylvain D; Rappsilber, Juri; Ricard-Blum, Sylvie; Cesareni, Gianni; Salwinski, Lukasz; Orchard, Sandra; Porras, Pablo; Panneerselvam, Kalpana; Hermjakob, Henning title: The IntAct database: efficient access to fine-grained molecular interaction data date: 2021-11-11 journal: Nucleic Acids Res DOI: 10.1093/nar/gkab1006 sha: 8bcb370708577d90bbdb5fdab473a59dd331c484 doc_id: 838456 cord_uid: tqv9kbb2 The IntAct molecular interaction database (https://www.ebi.ac.uk/intact) is a curated resource of molecular interactions, derived from the scientific literature and from direct data depositions. As of August 2021, IntAct provides more than one million binary interactions, curated by twelve global partners of the International Molecular Exchange consortium, for which the IntAct database provides a shared curation and dissemination platform. The IMEx curation policy has always emphasised a fine-grained data and curation model, aiming to capture the relevant experimental detail essential for the interpretation of the provided molecular interaction data. Here, we present recent curation focus and progress, as well as a completely redeveloped website which presents IntAct data in a much more user-friendly and detailed way. Biomolecular interactions are the fabric underlying almost all processes in living organisms, and they are determined by a broad array of experimental approaches, from focussed studies of pairwise interactions to large-scale determination of 10 000s of interactions in standardised high throughput experiments. However, observed molecular interactions are highly dependent on the biological and experimental conditions under which they are determined. Cellular systems, experimental protein tags sequence modifications, and experimental approaches all heavily influence the observed interaction. Since its inception in 2005, members of the International Molecular Exchange Consortium (IMEx) (1) have collaboratively curated molecular interaction data from the scientific literature and from direct data depositions, emphasizing a deep curation model aiming to capture interaction reports in sufficient detail to support subsequent comprehensive data presentation, aggregation, and analysis. In 2017, the IMEx Consortium became an ELIXIR core data resource (2) , recognising it as part of the fundamental infrastructure for life sciences. For an in-depth review of the current IMEx data model, curation strategies and collaborations, see (3) . The IntAct database of molecular interactions is used by all currently active IMEx partners (In-tAct, DIP (4), UniProt (5), MINT (6), MatrixDB (7), UCL ICS, IID (8)) as a common curation platform, and also acts as a common data dissemination platform, in parallel to the partners' own websites. While the detailed IMEx interaction data has always been available through download in the feature-rich PSI-MI XML format (9,10), many annotation details were not conveniently accessible through * To whom correspondence should be addressed. Tel: +44 1223 494671; Email: hhe@ebi.ac.uk C The Author(s) 2021. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. the IntAct website, and often users are not aware of the depth of available annotations. We are increasingly addressing this issue through the release of targeted datasets, in particular for sequence variations impacting interactions, and through a completely redeveloped website, which provides comprehensive filter and display tools to make optimal use of the rich annotation available in the IntAct database. Since the last IntAct NAR publication (11) , data content has grown from 408 000 (Jamuary 2014) to 1 114 500 (June 2021)) interaction evidences, and the number of referenced publications has risen from 12 500 to 22 500. This rapid increase is based on the integration of previously curated data from IMEx partners, as well as the ongoing curation work. The faster rise in interaction numbers compared to publication numbers reflects the increasing trend towards largescale interaction studies. In the same period, interactions from 21 publications have been retracted, usually due to retraction of the supporting publication. Several new datasets have been released, including two key collections: the 'Mutations dataset' (6) and the 'Coronavirus interactome' (7) . This dataset contains annotations describing the effect of small sequence changes on protein interactions. Captured changes comprise both natural variants and experimentally introduced sequence changes. This dataset is continuously maintained and updated, and since the original publication in February 2019 (12) it has grown from 28 000 to 72 000 mutation annotations. In order to fully reflect the importance of this data and to improve accessibility to it for users, in addition to web interface changes (see below), we have also introduced a dedicated tab-delimited download file format (https://www.ebi.ac.uk/ intact/download/datasets#mutations). After the outbreak of the COVID19 pandemic in Europe in early March 2020, we initiated an IMEx-wide initiative to record molecular interaction data related to SARS-CoV-2 and other members of the Coronaviridae family of viruses, along with human protein interactions of potential relevance for the disease's aethiopathology. Since its publication in November 2020 (13) , the Coronavirus interactome dataset has grown from 4400 interaction evidences derived from 151 publications to 9100 interaction evidences from 332 publications in June 2021, and is accessible at https://www.ebi.ac.uk/intact/resources/ datasets#coronavirus. Work is still actively ongoing to capture novel interactions, and details of known interactions such as the effects of variants, to further enhance this dataset. Given the fast pace at which COVID19-related data has been generated, the IMEx Consortium decided to allow the curation of preprints when the scientific interest contained in these publications justifies it. We will periodically review and update these datasets to ensure only data from peerreviewed publications is maintained in the database longterm. More information about IMEx's curation policy regarding preprints is provided at www.imexconsortium.org/ curation/. Curation practices and controlled vocabularies/ontologies are continuously updated, driven by the development of new methods like BioID (14) (term in PSI-MI Ontology https://www.ebi.ac.uk/ols/ontologies/mi/terms?iri = http% 3A%2F%2Fpurl.obolibrary.org%2Fobo%2FMI 1314). The IntAct sibling resource, the Complex Portal (15), now provides a reference resource for biomolecular complexes, and we are annotating complexes with Complex Portal identifiers as interacting objects where possible, in addition to interactions of proteins with small molecules, nucleic acids and polysaccharides such as glycosaminoglycans. We have redeveloped the IntAct web site (https://www.ebi. ac.uk/intact) to provide efficient, user-friendly access to In-tAct data content, with a focus on filter and display functionality to make the detailed interaction data accessible and useful through the user interface. The quick search provides autocompletion to facilitate selection of molecules of interest based on gene names, protein names, and accession numbers. The batch search supports multiple simultaneous query terms and subsequent result refinement. Results are shown both graphically and in tabular format, can be modified through comprehensive filter and visualisation options, and exported in both tabular and graphical formats. Figure 1 provides a view of the new IntAct web interface and its functionality, Figure 2 demonstrates the high level of detail provided for a single interaction. In addition to queries, species-specific interactomes ( Figure 3 ) and datasets like 'Alzheimers' (16) are available from tiles on the home page. The new IntAct public instance is deployed on the EMBL-EBI cloud using Kubernetes to manage the different containerized applications (the images have been built with Docker). The IntAct public interface is based on a Neo4j graph database and Apache Solr to enable the search and navigation features. An externally accessible API (https://www.ebi.ac.uk/intact/documentation/ technical corner#apis), developed in Java™ with the Spring framework to ease the implementation of the microservices architecture, serves data to both the web application and the Cytoscape app (17) . The web frontend is a single page application implemented with the Angular framework together with the EMBL-EBI Visual framework for general styling (https://www.ebi.ac.uk/style-lab/websites/). The network display (Figure 1 ) is based on Cytoscape.js (18) , the interaction detail view is the ComplexViewer (19) . The In-tAct web-based user interface has been specified using two rounds of user testing based on mockups. To provide a con- sistent user experience, we are co-ordinating the visualisation of interacting molecules in terms of shape (for molecule types) and default colour (for species) between the IntAct web interface and the IntAct Cytoscape app. Tissue specificity. Recent research emphasizes fundamental differences among cell type specific interactomes (20) . Detailed annotation of cell types/tissue has been standard practice in IMEx curation for a long time, but the information is currently partially in free text form and will benefit from standardisation and integration with ontologies like Experimental Factor Ontology (21), Brenda (22) , Uberon (23), Cell Line Ontology (24) , and Cellosaurus (25) . We are currently working on the restructuring of cell type/tissue annotation and increasing exposure of these data through download files and user interfaces. Rare diseases dataset. As part of our commitment to the clinical community, we are currently populating a rare disease dataset, with a focus on interactions affected by rare disease mutations. Approximately 5500 rare disease interactions have been annotated to date, assigning details such as kinetic parameters, variable experimental conditions or construct details, including binding surfaces and mutations that affect the interactions. The data features information about the amino acid changes, their effect over the inter- action and full reference to the experimental interaction evidence from which it was extracted. Currently, around 98% of the annotations are mapped to human proteins, providing high-quality experimental evidence of sequence change effects which directly relate to existing variation data. Credit attribution. The data presented here has been carefully curated over almost two decades by professional curators from twelve IMEx partners. To value scientific database curation as a key scientific activity in its own right, we are working on credit attribution for past and future IMEx curators through APICURON (26) and ORCID (https:// orcid.org/). IMEx: The International Molecular Exchange Consortium (IMEx), founded in 2005, is an international collaboration of twelve interaction data resources which coordinate their curation strategies. IntAct is an IMEx founding member, and provides the web-based curation platform used by all current IMEx partners. Interaction evidence: Interactions may have two or more participating molecules, and the number of observable interactors may depend on both biological and experimental constraints. As an example, the yeast-two-hybrid array technology (27) typically identifies only pairs of interactors (binary interactions), while techniques like tandem affinity purification (TAP) (28) and BioID (29) may identify two or more interacting molecules (n-ary interactions). Observed n-ary interactions are stored as such in the IntAct database, but for some download files and for visualisation, counting and comparison purposes, they are expanded into multiple binary interactions. In addition, one publication may use more than one experimental method to determine an interaction. One interaction evidence is one pair of interacting molecules, observed by one experimental approach, reported by one publication. In this manuscript, we use 'interaction' as a synonym for the technically more correct term of 'interaction evidence'. MI Score: The MI Score (30) is a quantitative estimate of the confidence in a given interaction. It is a normalized and D652 Nucleic Acids Research, 2022, Vol. 50, Database issue weighted count of independent interaction evidence and associated experimental methods. IntAct is open source, open data. The source code is available from https://github.com/intact-portal, all data is freely available through the web interface, API, and from https://www.ebi.ac.uk/intact/download under the CC BY 4.0 licence. Protein interaction data curation: the International Molecular Exchange (IMEx) consortium The ELIXIR Core Data Resources: fundamental infrastructure for the life sciences Towards a unified open access dataset of molecular interactions The Database of Interacting Proteins: 2004 update UniProt: the universal protein knowledgebase in 2021 MINT, the molecular interaction database: 2012 update MatrixDB: integration of new data with a focus on glycosaminoglycan interactions IID 2018 update: context-specific physical protein-protein interactions in human, model organisms and domesticated species Encompassing new use cases -level 3.0 of the HUPO-PSI format for molecular interactions Broadening the horizon-level 2.5 of the HUPO-PSI format for molecular interactions The MIntAct project-IntAct as a common curation platform for 11 molecular interaction databases Capturing variation impact on molecular interactions in the IMEx Consortium mutations data set The IMEx coronavirus interactome: an evolving map of Coronaviridae-host molecular interactions A promiscuous biotin ligase fusion protein identifies proximal and interacting proteins in mammalian cells Complex Portal 2018: extended content and enhanced visualization tools for macromolecular complexes A coordinated approach by public domain bioinformatics resources to aid the fight against Alzheimer's disease through expert curation of key protein targets IntAct App: a Cytoscape application for molecular interaction network visualisation and analysis Cytoscape.js: a graph theory library for visualisation and analysis ComplexViewer: visualization of curated macromolecular complexes Dual proteome-scale networks reveal cell-specific remodeling of the human interactome Modeling sample variables with an Experimental Factor Ontology BRENDA in 2019: a European ELIXIR core data resource Uberon, an integrative multi-species anatomy ontology CLO: The cell line ontology The cellosaurus, a cell-line knowledge resource APICURON: a database to credit and acknowledge the work of biocurators Interactome mapping provides a network of neurodegenerative disease proteins and uncovers widespread protein aggregation in affected brains Global landscape of protein complexes in the yeast Saccharomyces cerevisiae The necdin interactome: evaluating the effects of amino acid substitutions and cell stress using proximity-dependent biotinylation (BioID) and mass spectrometry Merging and scoring molecular interactions utilising existing community standards: tools, use-cases and a case study We would like to thank Andrés Baselga, Domitille Coq-Etchegaray, Théo Gauvrit, Maxime Lecomte, Carola Gómez Rodríguez and many other users who supported the development of the new website through systematic user testing and informal feedback.