This resource page brings together a curated set of human E3 ubiquitin ligases, with a focus on the catalytic components. To accommodate the wide diversity and complexity of E3 ligases, we use a multiscale approach, comparing E3 ligases across multiple hierarchies such as primary sequence, domain composition, 3D structure, and molecular function. These different, often orthogonal, layers are combined using an ML approach to group E3 ligases into meaningful families and subfamilies.
Contents
Subpages for each E3 ligase family, covering both catalytic and non-catalytic E3 components.
Each family page lists its member proteins.
Each protein links to a dedicated page that includes:
Basic Information
Classification
Protein sequence
Domain architecture
GO functional annotations
Enzyme–substrate interactions
Small molecule interactions
Database version: e3Ligome_202508
Curation and assembly of the human E3 Ligome
Integration of existing E3 ligase datasets showcases their overlaps and classification into key classes—RING, HECT, and RBR (including atypical and non-canonical catalytic mechanisms), defining the confidence scores for the human E3 ligome.
Metric learning for E3 ligases involves a linear combination of diverse molecular-level distance measures into a single emergent metric. The optimization procedure is based on the comparison of hierarchical clusters with the ground truth. A weakly supervised approach leverages orthogonal features of E3 ligases to reflect well-known E3 classes.
An unrooted hierarchical tree structure based on the optimized emergent distance metric captures authentic relationships and partitioning of E3 ligases into 13 families. Clusters reflect shared sequence, domain, structural, and functional features.
The functional landscape of the E3 ligome is visualized as a network of enriched GO annotation clusters corresponding to individual E3 families at all three ontologies (BP, MF, and CC). Individual nodes represent generic or specialized functions of E3 families.
Pairwise E3–substrate interactions, integrating known and predicted ESIs datasets along with direct and indirect protein–protein interactions. These map to approximately 75% of the ubiquitinated proteome, linking E3s to over 12,000 human proteins.
The Database, including all associated datasets, metadata, and documentation, is the intellectual property of the PROXIDRUGs consortium and the InnoDATA project.
The content of the Database is protected under applicable copyright and database protection laws.
Permitted Use
The Database is provided for research, academic, and non-commercial purposes only, unless explicitly stated otherwise.
Users are permitted to:
Access, download, and use the data.
Modify or build upon the data for academic, educational, or scientific purposes.
Share derived datasets, analyses, or visualizations with proper citation and explicit acknowledgment of “The Human E3 Ligome” project.
Prohibited Use
Use the Database or any derivative works for commercial purposes without prior written consent from the PROXIDRUGs consortia (InnoDATA PIs).
Redistribute or republish the Database (or substantial parts of it) without proper credit and adherence to the license terms.
Misrepresent the data, alter metadata to mislead, or use it in a way that violates ethical or scientific standards.
This license allows redistribution and adaptation, provided appropriate credit is given.
Contributions (e.g., data corrections, additions, or scripts) can be made through merge requests on GitLab.
Contributors must agree to license their contributions under the same CC BY 4.0 license.
Citation Requirements
If you use or refer to the Database in your publications, reports, or presentations, you must cite the paper:
Dutta et al. (2025, preprint)
Contact and Support
For inquiries, permissions, or support regarding the Database, please contact: Dr. Ramachandra M. Bhaskara, Goethe University Frankfurt.
Email: bhaskara@med.uni-frankfurt.de.
Website maintained by Alberto Cristiani, Arghya Dutta, and Siddhanta V. Nikte.
