PASSer:

Protein Allosteric Sites Server

Introduction

Allostery is the process by which proteins transmit perturbations caused by the binding effect at one site to another distal site. The allosteric process is fundamental in the regulation of activity. Identifying allosteric sites is important for allosteric drug development and has attracted a wide range of interests.

To approach this problem, PASSer is designed for accurate allosteric site prediction. PASSer provides three machine learning models: (1) ensemble learning through extreme gradient boosting and graph convolutional neural network; (2) automated machine learning through AutoGluon framework; and (3) learning to rank through lightgbm.

PASSer is deployed with trained machine learning models and has been extensively tested to complete prediction within seconds. The prediction submissions and result retrievals can be accessed via the web page or API.

Note: (1) All uploaded and analyzed data is deleted upon job completion. (2) The website is refreshed every day at 4 AM central time.

Website architecture

PASSer provides three well-trained machine learning models. NAR paper

Site prediction

Predictions can be finished within seconds. Submit jobs below. MLST paper

AutoML

Accurate prediction through automated machine learning. Frontiers paper

Tutorial

Check the tutorial with an interactive example. Tutorial page

Please input a protein identifier:

Please input a chain ID:

Choose a model:

Model	Execution Time	Type
Ensemble	Fast (1-2s)	Probability
AutoML	Slow (~20s)	Probability
Rank	Fast (1-2s)	Rank score

Please input a protein file: