Measurements Of Proteins’ Phase Separation Properties Based On Omic Experiment And Sequence Features

Eukaryotic cells contain numerous compartments that separate various biochemical reactions and regulatory processes from each other to ensure that biological processes are carried out properly and efficiently.Membraneless condensates,or membraneless organelles,are compartmentalized structures with different compositions,morphologies,and material states that mediate many important intracellular functions and can lead to disease development when abnormal.In recent years,there has been substantial evidence that liquid-liquid phase separation(LLPS)is an important mechanism driving the formation of membraneless condensates.The systematic identification of potential phase separation proteins and the measurement of their phase separation properties are fundamental to deepen the understanding of the mechanisms and biological functions behind the phase separation process.This not only depends on the assistance of computationl approach built on high-quality data,but also requires the development of high-throughput experimental method.In this study,based on published phase separation-related studies,we have integrated information of over 2,900 phase separation-related proteins supported by experimental evidence,performed phase separation-related sequence analysis and prediction,and developed an online accessible database,Pha Sep DB.We have developed a chromatin-binding protein capture method called Hi-MS,based on high-throughput chromosome conformation capture technology and mass spectrometry.We combined Hi-MS with 1,6-hexanediol treatment to disrupt phase separation condensates,defined an index to reflect the sensitivity of proteins to1,6-hexanediol treatment called AICAP(Anti-1,6-hexenidiol Index of Chromatin-Associated Proteins)which measures their phase separation properties.Based on Hi-MS and AICAP,we systematically measured the phase separation properties of more than 3000 proteins,and further found that proteins with dynamic regulatory functions and genomic higher-order structures were more sensitive to abnormal disruption of phase separation,while some proteins showed a more stable state.On the basis of Pha Sep DB,we constructed a dataset of phase separating proteins’ matterial states,and constructed Drop State,a classification model to distinguish the hydrogel/solid-like state of phase separated proteins based on protein sequence features.The classification efficacy(AUC)of Drop State was 0.725.Predicted hydrogel/solidlike state propensity coincided with the AICAP values obtained from Hi-MS experiments.Pha Sep DB is one of the first databases of phase separation-related proteins in the field,recording detail experimental and functional information.Hi-MS and its defined AICAP can systematically reflect the sensitivity of different proteins to phase separation disruption in physiological states,which is an effective complement to traditional experimental tools.The hydrogel/solid-like phase separation protein classification model Drop State predicts the state of matter of phase separated proteins by protein sequence features,and provides a reference for computational methods to reveal the biological function of phase separation from the perspective of material states.In summary,this study provides methods for measuring phase separation properties from the aspect of data support,omic experiment method and computational prediction.