Genetic architecture of cardiac dynamic flow volumes

Gomes, Bruna; Singh, Aditya; O’Sullivan, Jack W; Amar, David; Kostur, Mykhailo; Haddad, François; Salerno, Michael; Parikh, Victoria N.; Meder, Benjamin; Ashley, Euan A.

doi:10.1101/2022.10.05.22280733

Cited by 4 publications

(4 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…High-dimensional clinical data (HDCD; e.g., ECG, PPG, MRI) are valuable assets for clinical diagnosis, treatment, and prognosis, and provide a unique opportunity for studying the genetic basis of complex traits [1][2][3][4][5][6][7][8][9]. Recent technological progress in electronic health record (EHR) systems enables access to multiple HDCD modalities per individual [10][11][12].…”

Section: Introductionmentioning

confidence: 99%

Utilizing multimodal AI to improve genetic analyses of cardiovascular traits

Zhou,

Cosentino,

Yun

et al. 2024

Preprint

View full text Add to dashboard Cite

Electronic health record (EHR) and biobank datasets contain multiple high-dimensional clinical data (HDCD) modalities (e.g., ECG, Photoplethysmography (PPG), and MRI) for each individual. Access to multimodal HDCD provides a unique opportunity for genetic studies of complex traits because different modalities relevant to a single physiological system (e.g., circulatory system) encode complementary and overlapping information. We propose a novel multimodal deep learning method, M-REGLE, for discovering genetic associations from a joint representation of multiple complementary HDCD modalities. We showcase the effectiveness of this model by applying it to several cardiovascular modalities. M-REGLE jointly learns a lower-dimensional representation (i.e., latent factors) of multimodal HDCD using a convolutional variational autoencoder, performs genome-wide association studies (GWAS) on each latent factor, then combines the results to study the genetics of the underlying system. To validate the advantages of M-REGLE and multimodal learning, we apply it to common cardiovascular modalities (PPG and ECG), and compare its results to unimodal learning methods in which representations are learned from each data modality separately, but the downstream genetic analyses are performed on the combined unimodal representations. M-REGLE identifies 19.3% more loci on the 12-lead ECG dataset, 13.0% more loci on the ECG lead I + PPG dataset, and its genetic risk score significantly outperforms the unimodal risk score at predicting cardiac phenotypes, such as atrial fibrillation (Afib), in multiple biobanks.

show abstract

Section: Introductionmentioning

confidence: 99%

Utilizing multimodal AI to improve genetic analyses of cardiovascular traits

Zhou,

Cosentino,

Yun

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Recent advances in genomics research have enabled large-scale analyses assessing genetic contributions through methodologies including genome-wide association studies (GWAS), linkage disequilibrium score regression, Mendelian randomization, and various bioinformatic annotation approaches [12][13][14][15][16]. By integrating multiple genetic correlation and causality assessment techniques with functional genomic annotation, these cutting-edge analytic frameworks now allow comprehensive interrogation of potential biological mechanisms driving comorbidity from sequence variation through to gene regulatory effects [17][18][19].…”

Section: Introductionmentioning

confidence: 99%

Exploring the Common Genetic Underpinnings of Chronic Pulmonary Disease and Esophageal Carcinoma Susceptibility

Zhang,

Zhou,

et al. 2024

J. Cancer

View full text Add to dashboard Cite

Background: Pulmonary diseases and esophageal cancer are highly prevalent conditions with rising incidence worldwide. Prior evidence supports shared environmental and behavioral factors, but less is known regarding potential genetic links underlying this comorbidity. This study aimed to elucidate the complex genetic relationship between chronic lung diseases and esophageal cancer risk. Methods: Linkage disequilibrium score regression assessed the genetic correlation between esophageal cancer and asthma, COPD, and idiopathic pulmonary fibrosis leveraging extensive GWAS datasets. Pleiotropic analysis, gene-set enrichment, eQTL mapping, and mendelian randomization causality analyses were then conducted to identify specific shared genetic variants, enriched pathways, causal relationships and gene regulatory mechanisms connecting lung disease and cancer susceptibility. Results: Significant genetic correlations were observed between esophageal cancer and both COPD and asthma, but not idiopathic pulmonary fibrosis. Further analyses identified 13 pleiotropic loci and 6 shared genes including CHRNA4, ERBB3, and SMAD3, as well as pathways related to immune function. eQTL integration highlighted 53 genes like SOCS1, FGF2, and CHRNA5 with tissue-specific regulatory effects on disease risk. Bidirectional relationships were noted, whereby genetic predisposition to asthma and COPD increased esophageal cancer risk, while cancer liability reciprocally raised pulmonary fibrosis risk. Conclusions: These genomic analyses provide initial evidence that shared genetic factors may underpin the comorbidity between lung conditions and esophageal malignancy. The genes and pathways identified offer insights into biological mechanisms linking both diseases, aiding future screening, prevention and therapeutic efforts to mitigate this growing comorbidity burden.

show abstract

“…Narrow geometric phenotyping of the aorta neglects important three-dimensional aspects of its geometry, including elongation, tortuosity/unfolding, and curvature, all of which influence aortic hemodynamic function. Analyses of image-derived cardiac phenotypes acquired through the segmentation of cardiovascular magnetic resonance imaging have been used to uncover the genetic basis for cardiac structure and function 8,9 . Analyses of aortic 3D-structure in large community-based samples are required to uncover its demographic and clinical correlates.…”

Section: Introductionmentioning

confidence: 99%

Thoracic Aortic Three-Dimensional Geometry

Beeche,

Dib,

Zhao

et al. 2024

Preprint

View full text Add to dashboard Cite

Aortic structure and function impact cardiovascular health through multiple mechanisms. Aortic structural degeneration increases left ventricular afterload, pulse pressure and promotes target organ damage. Despite the impact of aortic structure on cardiovascular health, aortic 3D-geometry has yet to be comprehensively assessed. Using a convolutional neural network (U-Net) combined with morphological operations, we quantified aortic 3D-geometric phenotypes (AGPs) from 53,612 participants in the UK Biobank and 8,066 participants in the Penn Medicine Biobank. AGPs reflective of structural aortic degeneration, characterized by arch unfolding, descending aortic lengthening and luminal dilation exhibited cross-sectional associations with hypertension and cardiac diseases, and were predictive for new-onset hypertension, heart failure, cardiomyopathy, and atrial fibrillation. We identified 237 novel genetic loci associated with 3D-AGPs. Fibrillin-2 gene polymorphisms were identified as key determinants of aortic arch-3D structure. Mendelian randomization identified putative causal effects of aortic geometry on the risk of chronic kidney disease and stroke.

show abstract

Genetic architecture of cardiac dynamic flow volumes

Cited by 4 publications

References 55 publications

Utilizing multimodal AI to improve genetic analyses of cardiovascular traits

Utilizing multimodal AI to improve genetic analyses of cardiovascular traits

Exploring the Common Genetic Underpinnings of Chronic Pulmonary Disease and Esophageal Carcinoma Susceptibility

Thoracic Aortic Three-Dimensional Geometry

Contact Info

Product

Resources

About