SAN FRANCISCO — Applying machine learning to genome-wide association and electronic health record data may usher in a new era of precision in diagnosing and managing polycystic ovary syndrome.