Assess whether Monarch STRING ingest prioritizes high quality knowledge over computed results

[monicacecilia]

Background

The Monarch Knowledge Graph includes BioGRID as a second-order ingest from STRING. Results from interrogations to the Monarch KG provide users with information to follow interaction knowledge assertions back to STRING, where the primary sources STRING curates may be reviewed. ​​The interactions present in STRING include direct (physical) and indirect (functional) associations; they stem from computational prediction, knowledge transfer between organisms, and interactions aggregated from other (primary) databases. Pairwise experimental interaction evidence from BioGRID is integrated into STRING.

To assist Monarch users in assessing the nature of STRING edges in the Monarch graph, we now annotate Evidence and Conclusion Ontology codes on these edges. Please also note that Monarch only ingests relatively high-scoring STRING interaction pairs (combined score >0.7) into the Monarch graph. Scores are indicators of confidence, i.e., how likely STRING judges an interaction to be true, given the available evidence. Scores rank from 0 to 1, with 1 being the highest possible confidence. A score of 0.5 would indicate that roughly every second interaction might be erroneous (i.e., a false positive).

Action

To further reassure users, assess and precisely describe whether (and if so, how) the Monarch Initiative KG ingest process is indeed prioritizing high-quality curated knowledge over computed results – ensuring that good, high-quality curated BioGRID curation is favoured and captured within the established threshold.