Prediction of Osterici Radix Ingredients Related to Parkinson’s Disease Using Long Short-Term Memory(LSTM)
Year: 2024
Type: article
Abstract: The size of the global pharmaceutical industry market and the ratio of net sales to R&D investment is growing yearly.Still, as a high-cost-low-efficiency industry, the success rate is very low.Recently, pharmaceutical companies have been using computers or looking for new drug candidates from natural products to improve this.Medicinal herbs are natural products proven safe and effective and can be an important resource for developing new drugs However, medicinal herbs usually consist of dozens of ingredients, and due to the characteristics of multi-compound-multi-target, experimentally, it must be proved which ingredient is an active ingredient.Therefore, in this study, the ingredients related to Parkinson's disease were predicted using medicinal herbs using the machine learning model Long Short-Term Memory (LSTM).Medicinal herbs, which are highly related to Parkinson's disease symptoms, were selected through Dongui Bogam, and to predict the composition of medicinal herbs related to Parkinson's disease, the machine learning model LSTM learned the characteristics of compounds targeting G protein-coupled receptor (GPCR) related to Parkinson's disease.The components of Osterici Radix were applied to the machine learning model LSTM, and the top 30 compounds highly related to Parkinson's disease were predicted based on the probability.
Author: Myung-ku Lee
Institution: Korea Institute of Oriental Medicine
Language: English
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4404869469