Protein annotation and modelling servers at University College London

Area of Science:

• Bioinformatics
• Computational Biology
• Structural Biology

Background:

• Accurate protein structure and function prediction is crucial for biological research.
• Existing computational tools vary in accuracy and scope.
• The UCL Bioinformatics Group has developed a suite of algorithms to address these needs.

Purpose of the Study:

• To present the comprehensive suite of protein structure prediction and function annotation algorithms offered by the UCL Bioinformatics Group.
• To introduce BioSerf, an automated 3D modeling pipeline, highlighting its performance in CASP8.
• To provide researchers with accessible, high-quality computational tools via a web portal.

Main Methods:

• Utilizing established algorithms such as PSIPRED, pGenTHREADER, and DISOPRED2 for specific prediction tasks.
• Employing a fragment-assembly approach within the BioSerf pipeline for de novo protein modeling.
• Offering a web portal interface for user access to all developed tools.

Main Results:

• The portal offers tools for predicting secondary structure, protein fold, domains, transmembrane topology, metal binding sites, disorder, and function.
• BioSerf demonstrated strong performance in CASP8, ranking among the top five de novo modeling servers.
• A range of high-quality prediction algorithms are made available to the scientific community.

Conclusions:

• The UCL Bioinformatics Group web portal provides a valuable resource for protein structure and function analysis.
• The integrated suite of tools, including the BioSerf pipeline, enhances capabilities in computational protein research.
• These accessible web-based services support advancements in structural biology and drug discovery.