New Artificial Intelligence Model Could Speed Rare Disease Diagnosis
New Artificial Intelligence Model Could Speed Rare Disease Diagnosis
Publish Date: 2025-12-31 19:00:00
Source Domain: www.technology.org
- Harvard Medical School researchers developed an AI model called popEVE to identify disease-causing genetic variants among the vast number of genetic alterations in a human genome.
- popEVE scores each variant’s likelihood of causing disease based on cross-species evolutionary data and human population information.
- The model effectively distinguishes between benign and pathogenic variants, determines whether a variant leads to childhood or adult death, and identifies genetic variants connected with undiagnosed rare diseases.
- popEVE provided diagnoses in about one-third of cases with severe developmental disorders from a tested cohort, and identified previously unknown genetic causes for developmental disorders.
- The researchers aim to integrate popEVE into the clinical setting to offer quicker and more accurate diagnoses for rare genetic diseases, and potentially identify new drug targets for genetic conditions.
