Utilizing AI for Liver Cell Biology: Insights and Research Gaps through Analysis of the Human Protein Atlas (HPA) Liver Tissue Dataset

Abstract

This study investigates the use of artificial intelligence (AI) in liver cell biology by analyzing protein expression and localization patterns using the Human Protein Atlas (HPA) Liver Tissue Section dataset. Convolutional neural networks (CNNs) and multi-layer perceptron (MLP) models were employed to classify protein localization and predict expression levels, respectively. The CNN model achieved high test accuracy (87%) with balanced precision and recall, demonstrating strong performance in distinguishing cellular localization. The MLP model also achieved reliable predictions with a mean absolute error (MAE) of 0.14 on the test set. These findings highlight AI’s potential to advance liver-specific protein analysis, offering valuable insights for future research in liver biology and disease diagnosis. Future work could expand this framework to incorporate hybrid models for enhanced interpretability and accuracy.