PON-All predicts the tolerance (pathogenicity) of amino acid substitutions in multiple species. It is a machine learning-based approach and utilizes amino acid features, Gene Ontology (GO) annotations, evolutionary conservation, and annotations of functional sites. PON-All estimates the reliability of predictions and groups the variants into pathogenic, neutral and unknown classes. The method is based on a gradient boosting algorithm and it has been trained on a large data set. The method is fast and has high performance. Performance of PON-All has been extensively tested. For details, see here. It can also be used to predict variations for proteins from any organism.
PON-All was developed in collaboration between the groups of Prof. Mauno Vihinen in Lund University and Assoc. Prof. Yang Yang in Soochow University. You can also visit the mirror website in Soochow University, China.
There are three input formats. Identifier submission, Sequence submission or Genomic submission (only Human).
45573 variants were used for method training and 5360 variants were used for blind test. The data sets are freely available.
Result are displayed on the Results page. You can also provide your email, then the results are by email.
YANG Y, SHAO A, VIHINEN M. PON-All: Amino Acid Substitution Tolerance Predictor for All Organisms [J]. Frontiers in Molecular Biosciences, 2022, 9.(https://www.frontiersin.org/articles/10.3389/fmolb.2022.867572/full)