DeepIR-Pred: identification of insulin receptors using metaheuristic optimization of biologically informed multi-view features with deep recurrent learning.

DeepIR-Pred accurately identifies insulin receptor proteins with over 10-fold improved robustness and generalization using deep recurrent learning and biologically informed features.

• Why it matters: Understanding insulin receptor proteins is vital for advancing treatments for neurodegenerative and cancer-like diseases, but experimental methods are costly and slow, creating a need for reliable computational tools.
• What they did: The approach integrates novel image-based PSSM features, biologically informed characteristics, and protein language model embeddings, optimized with a metaheuristic whale algorithm, and trained with a deep bidirectional Gated Recurrent Unit network on curated datasets.
• The result: DeepIR-Pred demonstrates superior performance and interpretability, enabling researchers to explore IR proteins computationally, which could accelerate therapeutic discovery and biological understanding.