ProPythia: A Python Automated Platform for the Classification of Proteins Using Machine Learning

Abstract

A challenging problem in Bioinformatics is to predict protein structure, properties, activities or interactions from their aminoacid sequences. Sequence-derived physicochemical features of proteins have been used to support the development of Machine Learning (ML) models. However, tools and platforms to calculate features from protein sequences and train ML models are scarce and have limitations in terms of performance, user-friendliness and domains of application.

Here, a generic modular semi-automated platform for the classification of proteins based on their physicochemical properties using ML is proposed. The tool, developed as a Python package, facilitates the major tasks of ML and includes modules to read and alter sequences, calculate several types of protein descriptors, pre-process datasets, execute feature selection and dimensionality reduction, perform clustering, train and optimize ML models and make predictions with 8 different algorithms. ProPythia has an adaptable modular architecture being a versatile and easy-to-use tool to apply ML analysis over protein sequences. This platform was tested in the classification of membrane active anticancer and antimicrobial peptides. The package, its source code and documentation, including an user guide and case studies freely available at https://github.com/BioSystemsUM/propythia, it can also be installed through ‘pip install propythia’.