Predict isoform interaction conservation of a protein of interest with user submitted interaction partners


This tool predicts whether human protein-protein interactions (PPIs) are conserved between the alternative isoforms of the interacting proteins. Querying a protein using the corresponding UniProt ID or gene name will use experimental PPI data from the HI-II-14 and IntAct datasets together with domain annotations to predict the PPI conservation between the alternative isoforms. The advanced option allows the user to submit their own PPI data and view our domain-based predictions of the conservation of those interactions between the alternative isoforms.


Given an experimentally determined protein-protein interaction (PPI), our computational method uses the domain annotations of the proteins to map domain-domain interactions (DDIs) onto the PPI. The conservation of the PPI between the alternative isoforms of the proteins is predicted by whether the domains corresponding to the mapped DDI are present in the isoforms. If the interacting domain is not present in the alterntive isoform, the interaction is predicted to be lost, if the domain is present in the alternative isoform, the interaction is predicted to be conserved. In the case where multiple DDIs are mapped to a single PPI, the interaction is predicted to be lost only if domains corresponding to all DDI annotations are lost.


Protin-Protein Interactions HI-II-14 and IntAct
Domains Pfam
Domain-Domain Interactions 3did and DOMINE
Isoform Sequences UniProt

HI-II-14 and IntAct reference and predicted isoform interactomes
Reference Interactome Isoform interactome
Proteins Interactions Proteins Interactions
Reference Alternative Reference Predicted retained Predicted lost
HI-II-14 4,303 13,427 917 1,227 901 2,185 303
IntAct 9,023 29,775 2,944 4,471 4,363 12,651 1,709


Department of Bioengineering
Faculty of Engineering
McGill University
Montreal, Quebec H3A 0C3, Canada

Email: brandon.xia@mcgill.ca

Site built by: Mohamed Ghadie and Luke Lambourne