CLAP: a web-server for automatic classification of proteins with special reference to multi-domain proteins

Area of Science:

• Bioinformatics
• Computational Biology
• Structural Biology

Background:

• Protein function is best understood through structure, but sequence data is more abundant.
• Traditional alignment-based clustering methods struggle with multi-domain proteins due to sequence variability.
• Alignment-free methods are crucial for accurate protein classification, especially for complex protein architectures.

Purpose of the Study:

• To develop an alignment-free tool for classifying protein sequences.
• To overcome the limitations of alignment-based methods in handling multi-domain proteins.
• To leverage full-length protein sequence information for improved functional and architectural clustering.

Main Methods:

• Developed CLAP (Classification of Proteins), a web-server utilizing an alignment-free approach.
• Employed a pattern-matching algorithm to assign local matching scores (LMS) between sequences.
• Utilized full-length protein sequences, bypassing the need for domain definitions or alignments.

Main Results:

• CLAP successfully classifies protein sequences automatically.
• Pilot studies on protein kinases and immunoglobulins demonstrated high functional and domain architectural similarity within clusters.
• Statistically determined cut-offs in CLAP analysis produced clusters consistent with sub-family classifications.

Conclusions:

• CLAP is a versatile protein clustering tool effective regardless of domain assignment, order, length, or diversity.
• The method generates functionally relevant clusters with high domain architectural homogeneity for any protein sequence set.
• The CLAP web server is publicly accessible for academic research.