Evaluating the Traditional Chinese Medicine (TCM) Officially Recommended in China for COVID-19 Using Ontology-Based Side-Effect Prediction Framework (OSPF) and Deep Learning

Abstract

Ethnopharmacological relevance: Novel coronavirus disease (COVID-19) outbroke in Wuhan has imposed a huge influence onto the society in term of the public heath and economy. However, so far, no effective drugs or vaccines have been developed. Whereas, the Traditional Chinese Medicine (TCM) has been considered as a promising supplementary treatment for the disease owing to its clinically proven performance on many diseases even like severe acute respiratory syndrome (SARS). Meanwhile, many side-effect (SE) reports suggest the SE of the TCM prescriptions cannot be ignored in curing the COVID-19, especially because COVID-19 always simultaneously leads to dramatic degradation of the patients’ physical condition. How to evaluate the TCM regarding to their latent SE is a urgent challenge. Aim of the study: In this study, we use an ontology-based side-effect prediction framework (OSPF) developed in our previous work and Artificial Neural Network (ANN)-based deep learning to evaluate the TCM prescriptions that are officially recommended in China for novel coronavirus (COVID-19). Materials and methods: Firstly, we adopted the OSPF developed in our previous work, where an ontology-based model separate all the ingredients in a TCM prescription into two categories: hot and cold. Then, we established a database by converting each TCM prescription into a vector containing the ingredient dosage and the according hot/cold attribution as well as the safe/unsafe label. And, we trained the ANN model using this database, after which a safety indicator (SI), as the complementary percentage of side-effect (SE) possibility, is then given for each TCM prescription. According to the proposed SI from high to low, we re-organize the recommended prescription list. Secondly, by using this method, we also evaluate the safety indicators of some other famous TCM prescriptions that are not in the recommended list but are used traditionally to cure flu-like diseases for extending the potential treatments. Results: Based on the SI generated in the ANN model, FTS, PMSP, and SF are the safest ones in recommended list, which all own a more-than-0.8 SI. Whereas, JHQG, LHQW, SFJD, XBJ, and SHL are the prescriptions that are most likely unsafe, where the indicators are all below 0.2. In the extra list, the indicators of XC, XQRS, CC, and CHBX are all above 0.8, and at the meantime, XZXS, SJ, QW, and KBD’s indicators are all below 0.2. Conclusions: In total, there are seven TCM prescriptions which own the indicators more than 0.8, suggesting these prescriptions should be considered firstly in curing COVID-19, if suitable. We believe this work will provide a reasonable suggestion for the society to choose proper TCM as the supplementary treatment for COVID-19. Besides, this work also introduces a pilot and enlightening method for creating a more reasonable recommendation list of TCM to other diseases.

COVID‐19: Phylogenetic approaches may help in finding resources for natural cure

“…[NCCIH], 2020). Among them, Traditional Chinese Medicine (TCM) is on the top (Luo et al, 2020;Wang, Li, Yan, & Yao, 2020;Xu et al, 2020). TCMs have been used for thousands of years for the ailments of different diseases (Luo, Wang, Hesse-Fong, Lin, & Yuan, 2019;Ryu, 2018).…”

Artificial Intelligence in the Fight Against COVID-19: Scoping Review

“…In 31 studies , AI was used for diagnosing COVID-19 cases or identifying suspected COVID-19 cases based on various indicators, including computed tomography (CT) images [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38], x-ray images [39][40][41][42][43][44][45][46][47][48][49][50], laboratory tests [51,52], genome sequences [53], and respiratory patterns [54]. In 20 studies [55][56][57][58][59][60][61][62][63][64][65][66][67][68][69][70][71][72][73]…”

“…Specifically, 9 studies used AI for discovering drugs suitable for combating COVID-19 [55][56][57][58][59][60][61][62][63], and 2 studies used AI for repurposing commercially available drugs that could be used for treating COVID-19 [64,65]. There was 1 study that employed AI to predict the safety of using traditional Chinese medicine for COVID-19 [66]. In 4 studies [67][68][69][70], AI was used for discovering COVID-19 vaccines.…”

Going by the Numbers : Learning and Modeling COVID-19 Disease Dynamics

“…Wang et al [30] have proposed an “Ontology-based Side-effect Prediction Framework (OSPF)” for evaluating the “Traditional Chinese Medicine (TCM)” model using deep learning. They have trained an Artificial Neural Network (ANN) model with an architecture composed of three hidden layers on the TCM prescriptions used for the treatment of COVID-19 patients.…”