Genomic AI misses complex gene regulation, limiting clinical reliability
AI models trained on genomic sequences made accurate predictions about some DNA effects but missed complex gene regulation across tissues. Researchers found models struggle with long-range regulatory interactions and cellular context in human genomes from labs in multiple countries. That limitation slows clinical translation of AI genomic tools and raises caution for relying on black-box predictions.
Where this came from. One outlet reported this. We read it, then wrote the sixty words above.