Indexé dans
- Ouvrir la porte J
- Clés académiques
- RechercheRef
- Université Hamdard
- EBSCO AZ
- OCLC - WorldCat
- Publions
- Pub européen
- Google Scholar
Liens utiles
Partager cette page
Dépliant de journal
Revues en libre accès
- Agriculture et aquaculture
- Alimentation et nutrition
- Biochimie
- Bioinformatique et biologie des systèmes
- Business & Management
- Chimie
- Génétique et biologie moléculaire
- Immunologie & Microbiologie
- Ingénierie
- La science des matériaux
- Neurosciences & Psychologie
- Science générale
- Sciences cliniques
- Sciences environnementales
- Sciences médicales
- Sciences pharmaceutiques
- Sciences vétérinaires
- Soins infirmiers et soins de santé
Abstrait
Machine learning of 12-lead Electrocardiograms Identified Inherited Risk and Vulnerability to Atrial Fibrillation
Avtar Singh
Artificial intelligence (AI) models applied to 12-lead ECG waveforms can predict atrial fibrillation (AF), a hereditary and pathological arrhythmia. We hypothesized that the ECGAI-based risk estimation may have a genetic basis. We applied the ECGAI model to predict atrial fibrillation in ECGs from 39,986 UK bio bank participants without atrial fibrillation. Next, we performed a Genome-Wide Association Study (GWAS) of predicted atrial fibrillation risk. Three signals (P <5E8) were identified at the established AF sensitivity loci marked by the sarcomere gene TTN and the sodium channel genes SCN5A and SCN10A. We also identified two new loci near the VGLL2 and EXT1 genes. In contrast, GWAS of risk estimation from clinical variable models revealed different genetic profiles. The predicted AF risk from the EKGAI model is affected by sarcomere, ion channels, and genetic variation that suggests ascending pathways. The ECGAI model can identify individuals at risk of disease through specific biological pathways.