ECMPride: prediction of human extracellular matrix proteins based on the ideal dataset using hybrid features with domain evidence

Binghui Liu; Ling Leng; Xuer Sun; Yunfang Wang; Jie Ma; Yunping Zhu

doi:10.7717/peerj.9066

ECMPride: prediction of human extracellular matrix proteins based on the ideal dataset using hybrid features with domain evidence

PeerJ. 2020 Apr 29:8:e9066. doi: 10.7717/peerj.9066. eCollection 2020.

Authors

Binghui Liu^#¹, Ling Leng^#², Xuer Sun³, Yunfang Wang³, Jie Ma¹, Yunping Zhu^{1

4}

Affiliations

¹ State Key Laboratory of Proteomics, Beijing Proteome Research Center, National Center for Protein Sciences (Beijing), Beijing Institute of Life Omics, Beijing, China.
² Department of Central Laboratory, Peking Union Medical College Hospital, Peking Union Medical College and Chinese Academy of Medical Sciences, Beijing, China.
³ Tissue Engineering Lab, Institute of Health Service and Transfusion Medicine, Beijing, China.
⁴ Basic Medical School, Anhui Medical University, Anhui, China.

^# Contributed equally.

Abstract

Extracellular matrix (ECM) proteins play an essential role in various biological processes in multicellular organisms, and their abnormal regulation can lead to many diseases. For large-scale ECM protein identification, especially through proteomic-based techniques, a theoretical reference database of ECM proteins is required. In this study, based on the experimentally verified ECM datasets and by the integration of protein domain features and a machine learning model, we developed ECMPride, a flexible and scalable tool for predicting ECM proteins. ECMPride achieved excellent performance in predicting ECM proteins, with appropriate balanced accuracy and sensitivity, and the performance of ECMPride was shown to be superior to the previously developed tool. A new theoretical dataset of human ECM components was also established by applying ECMPride to all human entries in the SwissProt database, containing a significant number of putative ECM proteins as well as the abundant biological annotations. This dataset might serve as a valuable reference resource for ECM protein identification.

Keywords: Extracellular matrix proteins; Prediction tool; Proteomics; Random forest; Under-sampling ensemble method.

Grants and funding

This work was supported by the National Key Research Program of China (No. 2016YFB0201702, No. 2016YFC0901601, No. 2017YFC0906602, and No. 2017YFA0505002). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.