Skip to content

raghavagps/AHTpin

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 

Repository files navigation

AHTpin: Antihypertensive Peptide Prediction Platform

Welcome to the official repository for AHTpin, an in silico platform developed for predicting, screening, and designing antihypertensive peptides (AHTPs). The platform uses machine learning and QSAR-based approaches to identify bioactive peptides with potential antihypertensive activity.

Web Server: https://webs.iiitd.edu.in/raghava/ahtpin/ ZENODO : https://doi.org/10.5281/zenodo.20079810

Brief Description

Hypertension is one of the leading causes of cardiovascular diseases worldwide. Natural bioactive peptides have emerged as promising therapeutic agents due to their ability to reduce blood pressure with fewer side effects compared to synthetic drugs.

AHTpin was developed to provide a computational framework for identifying antihypertensive peptides using machine learning techniques. The platform integrates Support Vector Machine (SVM)-based regression and classification models trained on experimentally validated peptide datasets collected from public databases and scientific literature.

The system categorizes peptides into tiny, small, medium, and large peptide groups based on sequence length and applies specialized predictive models for each category. Various sequence-derived and chemical descriptors, including amino acid composition, atomic composition, and PaDEL molecular descriptors, were used to improve prediction performance.

In addition to prediction, AHTpin supports peptide screening, analog design, and mapping of antihypertensive regions within proteins, making it a valuable resource for peptide therapeutics, functional food research, and computational drug discovery.


Citation

Kumar, R., Chaudhary, K., Chauhan, J. S., Nagpal, G., Kumar, R., Sharma, M., & Raghava, G. P. S. (2015).
AHTpin: An in silico platform for predicting, screening and designing of antihypertensive peptides.
Scientific Reports, 5, 12512.
https://doi.org/10.1038/srep12512

About the Platform

AHTpin is a computational platform developed for identifying antihypertensive peptides from protein sequences and peptide libraries. The system integrates regression and classification machine learning models to predict peptide activity across different peptide lengths.

The platform categorizes peptides into:

  • Tiny peptides (Dipeptides & Tripeptides)
  • Small peptides (Tetrapeptides, Pentapeptides & Hexapeptides)
  • Medium peptides (Length 7–12)
  • Large peptides (Length >12)

The study compiled experimentally validated antihypertensive peptides from:

  • AHTPDB
  • BIOPEP
  • ACEpepDB
  • Published literature

Key Features

Machine Learning Models

  • SVM-based regression models
  • SVM-based classification models
  • QSAR-based peptide prediction

Prediction Modules

  • Antihypertensive peptide prediction
  • Peptide screening
  • Analog peptide design
  • Protein sequence mapping

Feature Extraction

  • Amino acid composition
  • Atomic composition
  • PaDEL chemical descriptors
  • G-scale physicochemical descriptors

Dataset Information

The platform was developed using:

  • 1745 experimentally validated antihypertensive peptides
  • Positive datasets from curated databases
  • Negative datasets generated from Swiss-Prot proteins

Peptide Categories:

  • 131 Dipeptides
  • 205 Tripeptides
  • 153 Tetrapeptides
  • 270 Pentapeptides
  • 199 Hexapeptides
  • 368 Medium peptides
  • 76 Large peptides

Performance Summary

Regression Models

  • Dipeptides: R = 0.701
  • Tripeptides: R = 0.543

Classification Models

  • Tetrapeptides: Accuracy = 76.67%
  • Pentapeptides: Accuracy = 72.04%
  • Hexapeptides: Accuracy = 77.39%
  • Medium peptides: Accuracy = 82.61%
  • Large peptides: Accuracy = 84.21%

Methodology

The prediction models were developed using:

  • Support Vector Machine (SVM)
  • Leave-One-Out Cross Validation (LOOCV)
  • External validation datasets

Descriptor selection was performed using:

  • Weka
  • RemoveUseless filtering
  • BestFirst feature selection
  • F-stepping optimization

Applications

  • Antihypertensive peptide discovery
  • Functional food research
  • Peptide therapeutic design
  • Bioinformatics and QSAR studies
  • Machine learning model development

Contact & Authors

Prof. Gajendra P. S. Raghava
raghava@iiitd.ac.in raghava@imtech.res.in

Developed at: Bioinformatics Centre,
CSIR-Institute of Microbial Technology, Chandigarh, India

License

This work is distributed under the
Creative Commons Attribution 4.0 International License (CC BY 4.0)

Acknowledgements

Supported by:

  • CSIR (Open Source Drug Discovery & GENESIS projects)
  • Department of Biotechnology (BTISNET), Government of India

We acknowledge all researchers and databases contributing to antihypertensive peptide research.

Source

Paper extracted from uploaded PDF. :contentReference[oaicite:0]{index=0}

About

AHTpin: An in silico platform for predicting, screening and designing of antihypertensive peptides

Topics

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors