AI Model and Experimental Study of Drug Effects on Planarians' Regeneration

Due to planarians' remarkable regenerative capabilities, they have been a popular biological model for researching the effects of drugs on tissue regeneration. These unique abilities provide insights into potential applications of regenerative medicine. Despite extensive research, machine-learning models have yet to be integrated into this field of study. This paper proposes the first artificial intelligence machine-learning models used to predict planarian morphology when exposed to different drugs and manipulations, leveraging datasets derived from 1716 experiments documented across 119 publications spanning from 1898 to 2018. Two machine learning models are explored in this paper: a random forest classifier model, and a deep learning neural network multi-output regression model. The initial results indicate that the random forest model achieved over 96% accuracy for predicting lethal, tail, and trunk morphologies, and 64%, 76%, and 80% accuracy for predicting pharynx, head, and eye morphologies. For the neural network regression model, the loss is at 0.18. Additionally, in order to understand drugs' effects on planarian morphology, learning, and memory, experimentation is conducted with different alcohol concentrations. This paper aims to facilitate the analysis of the effects of drugs on planarians using artificial intelligence technologies and to aid future experiments.