An Innovative Inducer of Platelet Production, Isochlorogenic Acid A, Is Uncovered through the Application of Deep Neural Networks.
Biomolecules
; 14(3)2024 Feb 23.
Article
em En
| MEDLINE
| ID: mdl-38540688
ABSTRACT
(1) Background:
Radiation-induced thrombocytopenia (RIT) often occurs in cancer patients undergoing radiation therapy, which can result in morbidity and even death. However, a notable deficiency exists in the availability of specific drugs designed for the treatment of RIT. (2)Methods:
In our pursuit of new drugs for RIT treatment, we employed three deep learning (DL) algorithms convolutional neural network (CNN), deep neural network (DNN), and a hybrid neural network that combines the computational characteristics of the two. These algorithms construct computational models that can screen compounds for drug activity by utilizing the distinct physicochemical properties of the molecules. The best model underwent testing using a set of 10 drugs endorsed by the US Food and Drug Administration (FDA) specifically for the treatment of thrombocytopenia. (3)Results:
The Hybrid CNN+DNN (HCD) model demonstrated the most effective predictive performance on the test dataset, achieving an accuracy of 98.3% and a precision of 97.0%. Both metrics surpassed the performance of the other models, and the model predicted that seven FDA drugs would exhibit activity. Isochlorogenic acid A, identified through screening the Chinese Pharmacopoeia Natural Product Library, was subsequently subjected to experimental verification. The results indicated a substantial enhancement in the differentiation and maturation of megakaryocytes (MKs), along with a notable increase in platelet production. (4)Conclusions:
This underscores the potential therapeutic efficacy of isochlorogenic acid A in addressing RIT.Palavras-chave
Texto completo:
1
Coleções:
01-internacional
Base de dados:
MEDLINE
Assunto principal:
Trombocitopenia
/
Ácido Clorogênico
/
Aprendizado Profundo
Limite:
Humans
País/Região como assunto:
America do norte
Idioma:
En
Ano de publicação:
2024
Tipo de documento:
Article