Open classification of de novo inorganic ion transport proteins

Abstract:

Transmembrane transport proteins play a crucial role in biological processes by selectively transporting substrates across membranes. As potential drug targets understanding their substrate specificity is essential in addressing various diseases. This study introduces TooT-Open-ICAT, a two-stage computational model for identifying inorganic ion transport proteins and predicting their specific substrates. The ICAT protein detector component effectively distinguishes inorganic ion transporters from other proteins, achieving a reliable performance with Matthews Correlation Coefficient (MCC) of 0.94. The specific substrate predictor demonstrates similar performance in predicting the substrates applying open classification to handle de novo transport proteins. The open classification approach, applied by thresholding network layers, shows promise in addressing unknown classes, emphasizing the importance of representative training datasets and threshold calculation for small classes. The method offers an analysis of 5 model organisms' transportomes holding promise for enhancing our understanding of inorganic ion transport proteins and their substrate.