Prediction of liquid-liquid phase separating proteins using machine learning

Abstract

Background: The liquid-liquid phase separation (LLPS) of biomolecules in cell underpins the formation of membraneless organelles, which are the condensates of protein, nucleic acid, or both, and play critical roles in cellular function. Dysregulation of LLPS is implicated in a number of diseases. Although the LLPS of biomolecules has been investigated intensively in recent years, the knowledge of the prevalence and distribution of phase separation proteins (PSPs) is still lag behind. Development of computational methods to predict PSPs is therefore of great importance for comprehensive understanding of the biological function of LLPS.

Results: Based on the PSPs collected in LLPSDB, we developed a sequence-based prediction tool for LLPS proteins (PSPredictor), which is an attempt at general purpose of PSP prediction that does not depend on specific protein types. Our method combines the componential and sequential information during the protein embedding stage, and, adopts the machine learning algorithm for final predicting. The proposed method achieves a tenfold cross-validation accuracy of 94.71%, and outperforms previously reported PSPs prediction tools. For further applications, we built a user-friendly PSPredictor web server ( http://www.pkumdl.cn/PSPredictor ), which is accessible for prediction of potential PSPs.

Conclusions: PSPredictor could identifie novel scaffold proteins for stress granules and predict PSPs candidates in the human genome for further study. For further applications, we built a user-friendly PSPredictor web server ( http://www.pkumdl.cn/PSPredictor ), which provides valuable information for potential PSPs recognition.

Keywords: Liquid–liquid phase separation (LLPS); Machine learning; Phase separation proteins (PSPs); Predictor.

Fig. 1

Relationship between percent recall and total percentage of human proteins accepted at given thresholds, for Model 0 and three best first generation prediction tools

Fig. 2

Fraction of proteins in each category (scaffold, regulator or client) predicted as PSPs by PSPredictor or PScore

Fig. 3

A Fraction of proteins in each tier group predicted as PSPs by first generation prediction tools and PSPredictor. B The number of predicted PSPs that overlapped between two prediction tools

Fig. 4

The model architectures of CBOW (A) and Skip-gram (B)

Fig. 5

2D vector projection of PSPs and non-PSPs by PCA