Pclust

Pclust constructs and analyzes all-against-all pairwise similarity networks (A2ApsN) to visualize protein relationships and infer functions of newly identified proteins by leveraging homologous links to experimentally characterized reference proteins using BLAST-derived similarity scores.

Key Features:

• All-against-all pairwise similarity network (A2ApsN): Represents pairwise sequence similarities between proteins as an A2ApsN for network-based analysis.
• BLAST-based similarity computation: Uses BLAST searches to compute pairwise similarities with emphasis on speed and efficiency.
• Avoidance of multiple sequence alignments: Relies on pairwise BLAST comparisons rather than multiple sequence alignments.
• Network visualization: Visualizes protein similarity networks to reveal homologous relationships and network context.
• 'Reference proteins' highlighting: Emphasizes experimentally studied and characterized "reference proteins" within the A2ApsN.
• Functional inference from homology: Infers protein function by transferring information from homologous and experimentally characterized proteins.
• Mitigation of annotation transfer errors: Uses emphasis on experimental reference proteins to reduce reliance on potentially erroneous automatic annotation transfers.

Scientific Applications:

• Function inference for novel proteins: Infers putative functions of newly identified proteins by network links to characterized homologs.
• Interpretation of experimental data: Places experimentally derived protein information into network context to aid functional interpretation.
• Annotation refinement and validation: Helps identify and reduce propagation of erroneous automatic annotations in sequence databases by leveraging experimental reference proteins.

Methodology:

Performs BLAST-based all-against-all pairwise similarity searches to build an A2ApsN, highlights experimentally characterized "reference proteins" within the network, and infers functions from homologous relationships without performing multiple sequence alignments.