Predict Enzyme Function with ActSeekN

ActSeekN predicts enzyme function from 3D protein structures using active-site structural similarity with the ActSeek algorithm.

Submit either PDB files or UniProt accession numbers. For UniProt inputs, ActSeekN fetches AlphaFold structures automatically when available.

Results include EC annotations, active-site residue mappings, and structure-based confidence metrics, and appear progressively in the query results page.

Input Limits and Processing

A maximum of 100 inputs are processed in a single query. If more entries are provided, the server removes invalid items and duplicates first, then processes only the first 100 cleaned inputs.

When a query contains multiple inputs, the server processes them asynchronously and the query results page updates progressively as individual results become available.

The query results page can be bookmarked and the browser window can be safely closed and revisited later.

Provide Input Data for Functional Annotation

Select one or more .pdb files (max 10 MB each).
Enter space- or comma-separated UniProt accessions (for example, P0A6L4 Q988B8).