Microbial interactions descriptions for humans and machines alike

Background: Microorganisms from multiple taxonomic domains engage in intermicrobial interactions to form communities in all biomes. Even with identical community members, these complex interactions can vary in time, space and due to changes in abiotic factors. The range of in silico, in vitro or in vivo methods used to study these interactions is equally wide, and all of them produce a large amount of data. However, there is no standard to report intermicrobial interactions and a dire need for an open access platform to datasets that support this evidence. Objectives: Improve findability and integration of microbial interactions data by humans and machines by enhancing their description via a set of standardized metadata. Methods: Microbial interaction metadata are specified using the LinkML modeling language to leverage validation capacity and interoperability with published ontologies or existing metadata. Durable hosting is enabled via interconnection to metadata in the knowledge base Wikidata embedded within a user-friendly interface and editing features for both humans and machines. Results: Metadata concerning the interactions themselves, associated community members, their environmental context and evidence methods were formalized by types, expected values, and ontology values where appropriate, to enable comparison between studies. We illustrate this metadata scheme by describing examples of microbial interactions within hosts. Once more microbial interactions are described accordingly, original hypotheses could be drawn and tested by bioinformaticians, network scientists and microbiologists