GTPbinder: Identification of GTP binding residue in protein sequences

Welcome to the official documentation for GTPbinder, a specialized computational tool developed for the prediction and identification of GTP-binding residues in protein sequences. GTP-binding proteins (G-proteins) are essential molecular switches in cellular processes, including signal transduction, protein synthesis, and intracellular trafficking.

Web Server: (https://webs.iiitd.edu.in/raghava/gtpbinder/)

Citation

Chauhan JS, Mishra NK, Raghava GPS (2010).Prediction of GTP interacting residues, dipeptides and tripeptides in a protein from its evolutionary information. BMC Bioinformatics, 11, 301.https://doi.org/10.1186/1471-2105-11-301

GitHub:-https://github.com/Manish-IIITD-repository/GTPbinder

About the Platform

GTPbinder was developed to assist researchers in identifying residues that interact with Guanosine Triphosphate (GTP). Since GTP-binding proteins often share low sequence similarity despite structural conservation, this platform utilizes evolutionary information in the form of Position-Specific Scoring Matrices (PSSM) to achieve high predictive accuracy.

The platform is designed to:

Annotate Proteomes: Identify potential GTP-binding sites in newly sequenced proteins .

Understand Mechanisms: Elucidate the structural and sequence requirements for GTP interaction .

Support Drug Design: Identify target residues for small-molecule inhibitors in signaling pathways .

Key Features

Predictive Modeling

Machine Learning: Built using Support Vector Machines (SVM) optimized with the RBF kernel .

Evolutionary Profiles: Utilizes PSI-BLAST generated PSSM profiles to capture residue conservation at interacting sites .

High Performance:

Single Sequence Mode: Achieved an accuracy of 70.81% and MCC of 0.42 .

PSSM Profile Mode: Achieved a significantly higher accuracy of 80.32% and MCC of 0.61 .

Validation: Evaluated using 5-fold cross-validation on a non-redundant dataset of 134 GTP-binding protein chains .

Integrated Features

Compositional Analysis: Includes models based on amino acid composition, dipeptide composition, and tripeptide composition .

Structural Context: The method accounts for the local environment of the central residue in overlapping sequence windows.

Overview of Model Development

The training dataset was derived from the PDB (Protein Data Bank), selecting 134 non-redundant protein chains with a sequence identity cut-off of 40% [cite: 1]. Interacting residues were defined as those with at least one atom within 3.5 Å of any atom of the GTP molecule [cite: 1].

Input Feature
Accuracy
MCC
Sensitivity
Specificity

Binary Profile
70.81%
0.42
70.32%
71.30%

PSSM Profile
80.32%
0.61
80.32%
80.32%

Applications

Signal Transduction: Analyzing G-protein coupled receptors (GPCRs) and small GTPases like Ras .

Protein Synthesis: Investigating elongation factors and initiation factors .

Traffic Control: Studying Rab proteins involved in vesicle docking and fusion .

Contact & Support

Prof. G.P.S. Raghava Head, Department of Computational BiologyIndraprastha Institute of Information Technology (IIIT-Delhi), India.Email: [e-mail]

License

This research and associated software are distributed under the Creative Commons Attribution License (CC BY 2.0), allowing for unrestricted use and distribution with proper credit to the original authors [cite: 1].