Welcome to PON-P3

PON-P3 predicts the tolerance (pathogenicity) of amino acid substitutions in human MANE specific proteins. It is a machine learning-based approach that utilizes gene and protein, variations and structural features. PON-P3 groups variants into pathogenic, neutral, and uncertain significance (VUS) classes. The method is based on a gradient boosting algorithm and has been trained on a large dataset. It is fast and has high performance. The performance of PON-P3 has been extensively tested and compared with state-of-the-art methods.

PON-P3 was developed in the group of Prof. Mauno Vihinen, Protein Structure and Bioinformatics Research group, Lund University, Sweden.

Get started

How to cite?

The PON-P3 manuscript has been submitted. In the meantime, cite the URL of the service.