A New Deep Learning and XAI-Based Algorithm for Features Selection in Genomics

Explainability improves medical AI reliability: The paper demonstrates that combining deep learning with explainable AI (XAI) methods helps identify which genes are actually driving diagnostic predictions, rather than treating the model as a "black box"—a critical requirement for clinicians to trust and validate AI recommendations in healthcare.

Feature selection reduces complexity without sacrificing accuracy: By using autoencoders to identify the most informative genes from large genomic datasets, the algorithm narrows down thousands of potential features to a meaningful subset, making results more interpretable and reducing the risk of overfitting to irrelevant patterns.

AI can accelerate precision medicine discovery: The application to Chronic Lymphocytic Leukemia shows that this approach can suggest specific genes for further medical investigation, potentially speeding up the discovery of biomarkers for personalized treatment—demonstrating practical value beyond academic benchmarks.

Responsible AI requires domain expertise integration: The paper's focus on identifying genes "meaningful for further medical investigation" highlights that responsible AI in genomics must work alongside domain experts rather than replacing clinical judgment, ensuring findings are biologically plausible and actionable.