Tools & Database Creation

SPAED - Segmentation of PhAge Endolysin Domains

SPAED is a specialized tool for identifying domains in phage endolysins using AlphaFold's Predicted Aligned Error (PAE) matrices. SPAED enables accurate domain boundary prediction by leveraging structural prediction confidence.

Key Features:

• Analyzes PAE files from AlphaFold2/3 or ColabFold to identify domain boundaries
• Hierarchical clustering approach for automatic domain segmentation
• High-throughput processing of multiple proteins
• Visualization tools for PyMOL to display predicted domains on structures
• Extraction of domain sequences in FASTA format

Web Interface:

SPAED is available through a user-friendly web interface at www.spaed.ca for quick analysis of individual proteins.

Citation:

Alexandre Boulay, Emma Cremelie, Clovis Galiez, Yves Briers, Elsa Rousseau, Roberto Vázquez. "SPAED: harnessing AlphaFold output for accurate segmentation of phage endolysin domains." Bioinformatics, Volume 41, Issue 10, October 2025, btaf531. https://doi.org/10.1093/bioinformatics/btaf531

---

InterPro

InterPro was used to get a domain annotation of the delineated domains (by SPAED). Only hits to Pfam with an E-value < 0.001 were considered for downstream analyses, but hits to other databases were included in PhaLP 2.0.

Citation:

Jones, P. et al. (2014). InterProScan 5: genome-scale protein function classification. Bioinformatics 30, 1236–1240.

SUBLYME - Software for Uncovering Bacteriophage LYsins in MEtagenomic datasets

SUBLYME is a machine learning tool specifically designed to identify bacteriophage lysins in metagenomic datasets. SUBLYME was trained using proteins from the PhaLP database and utilizes highly informative ProtT5 protein embeddings for accurate predictions.

Key Features:

• Machine learning-based prediction using ProtT5 protein embeddings
• Two-stage classification: lysin detection, then endolysin vs. VAL (Virion-Associated Lysin) classification
• Includes Prodigal for gene prediction from genomic sequences
• High-throughput processing with GPU support

Integration with EnVhog:

SUBLYME significantly expands the catalog of phage lytic proteins by mining the vast metagenomic resources available through the EnVhog database. This allows PhaLP 2.0 to include:

• Novel lysin candidates from environmental samples
• Lysins from uncultured phages
• Expanded diversity of protein families and domain architectures
• Proteins from diverse and extreme environments

Availability:

SUBLYME is available via PyPI (pip install sublyme) or as an Apptainer/Singularity container for complete genomic analysis pipelines.

---

MMseqs2 - Sequence Clustering

MMseqs2 (Many-against-Many sequence searching) is used in the SUBLYME pipeline for fast and sensitive sequence clustering and homology searching. MMseqs2 enables efficient processing of large metagenomic datasets by providing ultra-fast sequence comparisons while maintaining high sensitivity.

Citation:

Steinegger M and Soeding J. MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets. Nature Biotechnology, doi: 10.1038/nbt.3988 (2017).