GDReBase: A Knowledge Base for Relations between Human Gut Microbes and Diseases Based on Deep Learning

Abstract

Gut microbes play a prominent role in many aspects of human health, as seen through the increasing number of related studies. The accumulation of intestinal-flora-related studies enables us to better understand the various relationships between human gut microbes and other factors that affect the human body. However, the existing database does not meet the requirements of scientists to browse or retrieve the latest and most comprehensive published data. Thus, a knowledge base containing data related to gut microbes with updates occurring in real time would be highly valuable. We present a knowledge base of consistently curated relationships between human gut microbes and disease. By continuously and automatically collecting papers published in mainstream journals and using deep learning and NLP methods for entity relationship identification, GDReBase has now integrated 3674 diseases, 687 microbes, 7068 relationships, and 13,553 pieces of evidence from 518,286 papers, a figure that will continue to grow. GDReBase is a convenient and comprehensive resource for gut microbiology research and can be accessed free of charge.

Keywords

• gut microbes
• diseases
• knowledge database
• deep learning
• NLP