Deep learning enables genetic analysis of the human thoracic aorta

James P. Pirruccello; Mark D. Chaffin; Elizabeth L. Chou; Stephen J. Fleming; Honghuang Lin; Mahan Nekoui; Shaan Khurshid; Samuel F. Friedman; Alexander G. Bick; Alessandro Arduini; Lu Chen Weng; Seung Hoan Choi; Amer Denis Akkad; Puneet Batra; Nathan R. Tucker; Amelia W. Hall; Carolina Roselli; Emelia J. Benjamin; Shamsudheen K. Vellarikkal; Rajat M. Gupta; Christian M. Stegmann; Dejan Juric; James R. Stone; Ramachandran S. Vasan; Jennifer E. Ho; Udo Hoffmann; Steven A. Lubitz; Anthony A. Philippakis; Mark E. Lindsay; Patrick T. Ellinor

doi:10.1038/s41588-021-00962-4

Deep learning enables genetic analysis of the human thoracic aorta

James P. Pirruccello, Mark D. Chaffin, Elizabeth L. Chou, Stephen J. Fleming, Honghuang Lin, Mahan Nekoui, Shaan Khurshid, Samuel F. Friedman, Alexander G. Bick, Alessandro Arduini, Lu Chen Weng, Seung Hoan Choi, Amer Denis Akkad, Puneet Batra, Nathan R. Tucker, Amelia W. Hall, Carolina Roselli, Emelia J. Benjamin, Shamsudheen K. Vellarikkal, Rajat M. GuptaChristian M. Stegmann, Dejan Juric, James R. Stone, Ramachandran S. Vasan, Jennifer E. Ho, Udo Hoffmann, Steven A. Lubitz, Anthony A. Philippakis, Mark E. Lindsay, Patrick T. Ellinor

Research output: Contribution to journal › Article › peer-review

111 Scopus citations

Abstract

Enlargement or aneurysm of the aorta predisposes to dissection, an important cause of sudden death. We trained a deep learning model to evaluate the dimensions of the ascending and descending thoracic aorta in 4.6 million cardiac magnetic resonance images from the UK Biobank. We then conducted genome-wide association studies in 39,688 individuals, identifying 82 loci associated with ascending and 47 with descending thoracic aortic diameter, of which 14 loci overlapped. Transcriptome-wide analyses, rare-variant burden tests and human aortic single nucleus RNA sequencing prioritized genes including SVIL, which was strongly associated with descending aortic diameter. A polygenic score for ascending aortic diameter was associated with thoracic aortic aneurysm in 385,621 UK Biobank participants (hazard ratio = 1.43 per s.d., confidence interval 1.32–1.54, P = 3.3 × 10⁻²⁰). Our results illustrate the potential for rapidly defining quantitative traits with deep learning, an approach that can be broadly applied to biomedical images.

Original language	English (US)
Pages (from-to)	40-51
Number of pages	12
Journal	Nature Genetics
Volume	54
Issue number	1
DOIs	https://doi.org/10.1038/s41588-021-00962-4
State	Published - Jan 2022
Externally published	Yes

ASJC Scopus subject areas

Genetics

Access to Document

10.1038/s41588-021-00962-4

Cite this

Pirruccello, J. P., Chaffin, M. D., Chou, E. L., Fleming, S. J., Lin, H., Nekoui, M., Khurshid, S., Friedman, S. F., Bick, A. G., Arduini, A., Weng, L. C., Choi, S. H., Akkad, A. D., Batra, P., Tucker, N. R., Hall, A. W., Roselli, C., Benjamin, E. J., Vellarikkal, S. K., ... Ellinor, P. T. (2022). Deep learning enables genetic analysis of the human thoracic aorta. Nature Genetics, 54(1), 40-51. https://doi.org/10.1038/s41588-021-00962-4

Pirruccello, JP, Chaffin, MD, Chou, EL, Fleming, SJ, Lin, H, Nekoui, M, Khurshid, S, Friedman, SF, Bick, AG, Arduini, A, Weng, LC, Choi, SH, Akkad, AD, Batra, P, Tucker, NR, Hall, AW, Roselli, C, Benjamin, EJ, Vellarikkal, SK, Gupta, RM, Stegmann, CM, Juric, D, Stone, JR, Vasan, RS, Ho, JE, Hoffmann, U, Lubitz, SA, Philippakis, AA, Lindsay, ME & Ellinor, PT 2022, 'Deep learning enables genetic analysis of the human thoracic aorta', Nature Genetics, vol. 54, no. 1, pp. 40-51. https://doi.org/10.1038/s41588-021-00962-4

@article{52077d2405f4489e8c9d25747cd34c0d,

title = "Deep learning enables genetic analysis of the human thoracic aorta",

abstract = "Enlargement or aneurysm of the aorta predisposes to dissection, an important cause of sudden death. We trained a deep learning model to evaluate the dimensions of the ascending and descending thoracic aorta in 4.6 million cardiac magnetic resonance images from the UK Biobank. We then conducted genome-wide association studies in 39,688 individuals, identifying 82 loci associated with ascending and 47 with descending thoracic aortic diameter, of which 14 loci overlapped. Transcriptome-wide analyses, rare-variant burden tests and human aortic single nucleus RNA sequencing prioritized genes including SVIL, which was strongly associated with descending aortic diameter. A polygenic score for ascending aortic diameter was associated with thoracic aortic aneurysm in 385,621 UK Biobank participants (hazard ratio = 1.43 per s.d., confidence interval 1.32–1.54, P = 3.3 × 10−20). Our results illustrate the potential for rapidly defining quantitative traits with deep learning, an approach that can be broadly applied to biomedical images.",

author = "Pirruccello, \{James P.\} and Chaffin, \{Mark D.\} and Chou, \{Elizabeth L.\} and Fleming, \{Stephen J.\} and Honghuang Lin and Mahan Nekoui and Shaan Khurshid and Friedman, \{Samuel F.\} and Bick, \{Alexander G.\} and Alessandro Arduini and Weng, \{Lu Chen\} and Choi, \{Seung Hoan\} and Akkad, \{Amer Denis\} and Puneet Batra and Tucker, \{Nathan R.\} and Hall, \{Amelia W.\} and Carolina Roselli and Benjamin, \{Emelia J.\} and Vellarikkal, \{Shamsudheen K.\} and Gupta, \{Rajat M.\} and Stegmann, \{Christian M.\} and Dejan Juric and Stone, \{James R.\} and Vasan, \{Ramachandran S.\} and Ho, \{Jennifer E.\} and Udo Hoffmann and Lubitz, \{Steven A.\} and Philippakis, \{Anthony A.\} and Lindsay, \{Mark E.\} and Ellinor, \{Patrick T.\}",

note = "Publisher Copyright: {\textcopyright} 2021, The Author(s), under exclusive licence to Springer Nature America, Inc.",

year = "2022",

month = jan,

doi = "10.1038/s41588-021-00962-4",

language = "English (US)",

volume = "54",

pages = "40--51",

journal = "Nature Genetics",

issn = "1061-4036",

publisher = "Nature Research",

number = "1",

}

TY - JOUR

T1 - Deep learning enables genetic analysis of the human thoracic aorta

AU - Pirruccello, James P.

AU - Chaffin, Mark D.

AU - Chou, Elizabeth L.

AU - Fleming, Stephen J.

AU - Lin, Honghuang

AU - Nekoui, Mahan

AU - Khurshid, Shaan

AU - Friedman, Samuel F.

AU - Bick, Alexander G.

AU - Arduini, Alessandro

AU - Weng, Lu Chen

AU - Choi, Seung Hoan

AU - Akkad, Amer Denis

AU - Batra, Puneet

AU - Tucker, Nathan R.

AU - Hall, Amelia W.

AU - Roselli, Carolina

AU - Benjamin, Emelia J.

AU - Vellarikkal, Shamsudheen K.

AU - Gupta, Rajat M.

AU - Stegmann, Christian M.

AU - Juric, Dejan

AU - Stone, James R.

AU - Vasan, Ramachandran S.

AU - Ho, Jennifer E.

AU - Hoffmann, Udo

AU - Lubitz, Steven A.

AU - Philippakis, Anthony A.

AU - Lindsay, Mark E.

AU - Ellinor, Patrick T.

PY - 2022/1

Y1 - 2022/1

N2 - Enlargement or aneurysm of the aorta predisposes to dissection, an important cause of sudden death. We trained a deep learning model to evaluate the dimensions of the ascending and descending thoracic aorta in 4.6 million cardiac magnetic resonance images from the UK Biobank. We then conducted genome-wide association studies in 39,688 individuals, identifying 82 loci associated with ascending and 47 with descending thoracic aortic diameter, of which 14 loci overlapped. Transcriptome-wide analyses, rare-variant burden tests and human aortic single nucleus RNA sequencing prioritized genes including SVIL, which was strongly associated with descending aortic diameter. A polygenic score for ascending aortic diameter was associated with thoracic aortic aneurysm in 385,621 UK Biobank participants (hazard ratio = 1.43 per s.d., confidence interval 1.32–1.54, P = 3.3 × 10−20). Our results illustrate the potential for rapidly defining quantitative traits with deep learning, an approach that can be broadly applied to biomedical images.

AB - Enlargement or aneurysm of the aorta predisposes to dissection, an important cause of sudden death. We trained a deep learning model to evaluate the dimensions of the ascending and descending thoracic aorta in 4.6 million cardiac magnetic resonance images from the UK Biobank. We then conducted genome-wide association studies in 39,688 individuals, identifying 82 loci associated with ascending and 47 with descending thoracic aortic diameter, of which 14 loci overlapped. Transcriptome-wide analyses, rare-variant burden tests and human aortic single nucleus RNA sequencing prioritized genes including SVIL, which was strongly associated with descending aortic diameter. A polygenic score for ascending aortic diameter was associated with thoracic aortic aneurysm in 385,621 UK Biobank participants (hazard ratio = 1.43 per s.d., confidence interval 1.32–1.54, P = 3.3 × 10−20). Our results illustrate the potential for rapidly defining quantitative traits with deep learning, an approach that can be broadly applied to biomedical images.

UR - http://www.scopus.com/inward/record.url?scp=85119969043&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85119969043&partnerID=8YFLogxK

U2 - 10.1038/s41588-021-00962-4

DO - 10.1038/s41588-021-00962-4

M3 - Article

C2 - 34837083

AN - SCOPUS:85119969043

SN - 1061-4036

VL - 54

SP - 40

EP - 51

JO - Nature Genetics

JF - Nature Genetics

IS - 1

ER -

Deep learning enables genetic analysis of the human thoracic aorta

Abstract

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this