The first outcome of the IDPfun project are the scientific software tools developed for studying IDPs and inferring their properties. They include MobiDB-lite, which detects intrinsically disordered regions and predicts their structural properties solely from the protein sequence. Another tool is the PED web service, a validation pipeline for evaluating the predicted three-dimensional coordinates of IDP structural ensembles. RING is a software that calculates network properties and examines allosteric communication within an IDP ensemble. Additionally, AlphaFold-Disorder enables the inference of intrinsically disordered binding regions and short sequence motifs from predicted protein complexes, while FLIPPER and MOBI derive binding and disorder information from experimental structural data. Other software focus on analyzing IDP evolution and automatically extracting knowledge from the literature.
The second significant achievement of IDPfun pertains to the standardization and integration of IDP data into public databases while adopting best practices. The IDPfun consortium developed the Minimum Information About Disorder Experiments (MIADE) guidelines in collaboration with the IDP working group of the Human Proteome Organization - Proteomics Standards Initiative (HUPO-PSI) and the ELIXIR IDP Community. They also integrated the Evidence & Conclusion Ontology (ECO) with new terms for describing IDP experiments and created curation guidelines for community annotation of IDPs from the literature. The IDPfun consortium organized the Critical Assessment of protein Intrinsic Disorder (CAID) challenge, which evaluates IDP prediction methods and promotes standardization of IDP software. CAID is globally recognized, connected with ELIXIR services, and coordinated with the Critical Assessment of protein Structure Prediction (CASP). Also, IDPfun established ontological definitions for describing IDP functions. New terms were integrated into the Gene Ontology (GO), and IDPfun actively participated in the organization of the Critical Assessment of Function Annotation (CAFA), which evaluates function prediction methods.
The third major accomplishment of the IDPfun project involved the integration of IDP knowledge into public databases. New IDP models were incorporated into IDPfun curated databases such as DisProt, PED, ELM, and FuzDB. Additionally, the output of the developed software tools was integrated into the MobiDB aggregation database. By standardizing IDP function and the output format of prediction tools, data exchange between IDPfun resources and core resources maintained at the European Bioinformatics Institute (EBI), including InterPro, PDBe-KB, and UniProtKB, became possible. Moreover, the standardization of IDP ensemble structural data enabled the aggregation of PED records with experimental structural data from the Protein Data Bank (PDB) consortium in the 3D-beacons system, providing a centralized interface for accessing all publicly available structural data.
Scientific and technological achievements that contributed to the overall mission of the IDPfun research project were effectively disseminated through various activities aimed at raising awareness of IDPs in the life sciences community: integrating IDP knowledge into major databases, publishing in high-impact journals, incorporating IDP-related challenges into international bioinformatics assessments (CAFA and CASP), organizing conferences, online seminars, and hackathons, maintaining an active presence in social networks, establishing partnerships with other actions (PhasAGE, ML4NGP, etc.), and consortia (ELIXIR, GO, ECO, HUPO-PSI, InterPro, UniProtKB, PDB, etc.). The IDPfun project also provided training to researchers on utilizing IDP resources through protocol articles, webinars, training schools, and bootcamps.
