Researchers led by Michigan State University developed an AI model that predicts chemical effects on gene expression, speeding up drug discovery. The system analyses chemical structures to determine whether compounds increase or decrease activity in specific genes.

The model was trained on vast datasets of experimental results, allowing it to filter complex biological signals and produce reliable predictions. The approach allows virtual screening of millions of compounds, reducing the need for early-stage lab testing.

Study identified promising compounds for treating hepatocellular carcinoma and idiopathic pulmonary fibrosis, two diseases with limited therapeutic options. Lab and animal tests confirmed several compounds reduced tumour growth or showed promise for lung disease treatment.

Findings highlight the growing role of AI in medicine, with researchers emphasising that collaboration across computational science, biology, and clinical practice remains essential to bringing new therapies from discovery to real-world use.