Pcons.net

Pcons.net predicts protein tertiary structures and assesses model quality using a consensus meta-server that integrates outputs from PSI-BLAST-based modeling, fold-recognition servers, and structural quality estimators.

Key Features:

• Consensus approach: Combines predictions from multiple servers into a consensus score to improve reliability and accuracy of structural models.
• Model generation using PSI-BLAST and fold recognition: Produces models for PSI-BLAST-detectable targets and employs fold-recognition servers to identify distant structural homologs for challenging targets.
• Local and global quality estimation: Evaluates model quality with tools such as ProQ, ProQres and ProQprof, and uses Pcons-local for local quality prediction.
• Multi-source integration (Pcons5): Integrates consensus analysis, structural evaluation, and scores from fold-recognition servers to produce a refined consensus score.

Scientific Applications:

• Protein structure prediction: Generation of tertiary structure models for protein sequences using consensus and fold-recognition methods.
• Model quality assessment: Estimation of local and global model reliability using ProQ, ProQres, ProQprof, and Pcons-local.
• Detection of distant homologs: Identification of remote structural homologs via fold-recognition servers to enable modeling of difficult targets.
• Structural analysis for functional inference: Provision of structural models to support studies of protein function and interactions.

Methodology:

Sequence input is analyzed; for PSI-BLAST-detectable targets models are generated from sequence alignments, while fold-recognition servers are used to find distant homologs for other targets; results are evaluated with Pcons and ProQ (including ProQres and ProQprof) and integrated into a consensus, with Pcons-local and Pcons5 combining consensus analysis, structural evaluation, and fold-recognition scores.