CBDPS 1.0: A Python GUI Application for Machine Learning Models to Predict Bitter-Tasting Children's Oral Medicines.

Area of Science:

• Pharmacology
• Computational Chemistry
• Pediatrics

Background:

• Bitter taste perception is an innate aversive response, crucial for avoiding poisons.
• Children exhibit heightened sensitivity to bitter tastes, leading to poor adherence with medication.
• Effective taste-masking strategies are essential for improving pediatric drug compliance.

Purpose of the Study:

• To develop and validate a machine learning model for predicting the bitterness of pediatric medications.
• To create a user-friendly system for assessing drug bitterness based on chemical structure.
• To address the gap in existing models and databases concerning children's medication bitterness.

Main Methods:

• Trained four distinct machine learning models to predict bitterness.
• Developed the Children's Bitter Drug Prediction System (CBDPS) using Tkinter.
• Utilized the Simplified Molecular-Input Line-Entry System (SMILES) for chemical structure input.

Main Results:

• The XGBoost-Molecular ACCess System (XgBoost-MACCS) model achieved an 88% accuracy rate through cross-validation.
• The CBDPS system can predict the bitterness of single or multiple compounds.
• The developed model demonstrates potential for identifying bitter drug components.

Conclusions:

• The Children's Bitter Drug Prediction System (CBDPS) shows promise in predicting pediatric drug bitterness.
• This tool can aid in developing better-tasting medications for children.
• Further development could enhance drug compliance and palatability in pediatric populations.