# ECOD v294 Release Notes

**Release Date:** March 8, 2026

## Overview

ECOD v294 is a major update combining three workstreams: full Pfam 38.2 reconciliation, archaeal proteome expansion, and curator-driven consistency fixes. This release introduces 25,921 new protein families, adds 157,335 archaeal domains from 65 target classes, and incorporates boundary corrections and reclassifications across the hierarchy.

## Key Changes

### Pfam 38.2 Reconciliation

Complete bidirectional reconciliation against Pfam database version 38.2:
- **661,822 domain reassignments** based on updated Pfam HMM profiles
- **25,921 new F-groups** created from Pfam 38.2 families
- **1,168 F-groups deprecated** (emptied by reassignment)
- **273,016 domains demoted to .0** (no valid Pfam 38.2 hit)
- **21,389 domains promoted** from .0 to classified F-groups
- **97,153 marginal AFDB fragment domains** deprecated (48P set)

### Archaeal Proteome Expansion

First systematic incorporation of archaeal predicted structures:
- **157,335 new domains** from **98,146 proteins** across 65 archaeal species
- Sources: AlphaFold DB (96,651), Prodigal gene predictions (33,700), UniParc (26,984)
- **1,630 new F-groups** created via DPAM + Pfam annotation pipeline
- Domain assignment: 115,717 family-specific + 41,618 topology-only (.0)

### Consistency Curation

Curator-driven corrections improving classification accuracy:
- X-group merges and domain splits
- Domain boundary corrections
- Family reclassifications
- Beta propeller remediation
- Helicase domain fixes

## Statistics

### Domain Coverage

| Metric | Count |
|--------|-------|
| Total domains | 2,720,205 |
| PDB domains | 1,146,444 |
| AlphaFold/predicted domains | 1,573,761 |

### Classification Hierarchy

| Level | Count |
|-------|-------|
| X-groups (Possible Homology) | 2,459 |
| H-groups (Homology) | 3,719 |
| T-groups (Topology) | 3,954 |
| F-groups (Family) | 40,663 |

### Representatives

| Type | Count |
|------|-------|
| F40 clustering representatives | 464,955 |
| F70 clustering representatives | 889,259 |
| F99 clustering representatives | 1,491,147 |

## Distribution Files

All distribution files are available at: http://prodata.swmed.edu/ecod/distributions/

### Main Files
- `ecod.v294.domains.txt` - Complete domain list with classifications
- `ecod.v294.names.txt` - Simplified domain index
- `ecod.v294.fa` - FASTA sequences for all domains
- `ecod.v294.hierarchy.txt` - Classification hierarchy with domain counts
- `ecod.v294.f_id_pfam_acc.txt` - ECOD F-group to Pfam accession mapping

### Clustering Representatives
- `ecod.v294.F40.*` - 40% sequence identity representatives
- `ecod.v294.F70.*` - 70% sequence identity representatives
- `ecod.v294.F99.*` - 99% sequence identity representatives

### Structure Files
- `ecod.v294.F40.pdb.tar.gz` - F40 representative PDB structures (13GB)

### Search Databases
- `ecod.v294.F40.hhm_db.tar.gz` - HHsuite profile database (F40 representatives, 3.9GB)
- `ecod.v294.blast.*` - BLAST-formatted database (domain-level)
- `ecod.v294.chainwise.*` - BLAST-formatted database (chain-level)

### Checksums
- `ecod.v294.md5` - MD5 checksums for all files

## Changes from v293.1

| Metric | v293.1 | develop294 | Change |
|--------|--------|------------|--------|
| Total domains | 2,626,470 | 2,720,205 | +93,735 |
| PDB domains | 1,110,671 | 1,146,444 | +35,773 |
| Predicted domains | 1,515,799 | 1,573,761 | +57,962 |
| F-groups | 20,632 | 40,663 | +20,031 |
| X-groups | 2,455 | 2,459 | +4 |
| T-groups | 3,948 | 3,954 | +6 |

## Data Sources
- PDB structures current as of February 2026
- AlphaFold DB v4 (48 proteomes + SwissProt set)
- Archaeal proteomes from AlphaFold DB, Prodigal, and UniParc
- Pfam 38.2 HMM profiles
- UniProt 2024_06 annotations

## Citation

Please cite ECOD as:
> R. D. Schaeffer, K. E. Medvedev, A. Andreeva, S. R. Chuguransky, B. L. Pinto, J. Zhang, Q. Cong, A. Bateman, N. V. Grishin. (2025) ECOD: integrating classifications of protein domains from experimental and predicted structures. Nucleic Acids Res. 53(D1): D411-D418.

## Contact

For questions or issues, please contact: ecod.database@gmail.com