Skip to main content
NEJM AI homepage

Monthly & Annual Options

Subscribe Now

Monthly & Annual Options

Subscribe Now

This article is available to subscribers. Subscribe now

Already a subscriber? Sign in

AI-MARRVEL — A Knowledge-Driven AI System for Diagnosing Mendelian Disorders

Authors: Dongxue Mao, Ph.D. https://orcid.org/0000-0002-8443-2833, Chaozhong Liu, Ph.D. https://orcid.org/0000-0002-6960-1977, Linhua Wang, Ph.D. https://orcid.org/0000-0002-6717-860X, Rami AI-Ouran, Ph.D. https://orcid.org/0009-0004-6255-4727, Cole Deisseroth, B.S. https://orcid.org/0000-0001-9097-1617, Sasidhar Pasupuleti, M.S. https://orcid.org/0000-0003-0310-5812, Seon Young Kim, M.S. https://orcid.org/0009-0003-7542-545X, Lucian Li, B.S. https://orcid.org/0000-0002-9462-6527, Jill A. Rosenfeld, Ph.D. https://orcid.org/0000-0001-5664-7987, Linyan Meng, Ph.D. https://orcid.org/0000-0001-7474-9178, Lindsay C. Burrage, M.D., Ph.D. https://orcid.org/0000-0002-5108-8861, Michael F. Wangler, M.D. https://orcid.org/0000-0001-5245-5910, Shinya Yamamoto, D.V.M., Ph.D. https://orcid.org/0000-0003-2172-8036 on behalf of for Undiagnosed Diseases Network,* , Michael Santana, M.S. https://orcid.org/0009-0002-3209-1229, Victor Perez, B.S. https://orcid.org/0009-0006-6660-7907, Priyank Shukla, Ph.D. https://orcid.org/0009-0001-9308-1465, Christine M. Eng, M.D. https://orcid.org/0009-0008-8627-8228, Brendan Lee, M.D., Ph.D. https://orcid.org/0000-0001-8573-4211, Bo Yuan, Ph.D. https://orcid.org/0000-0001-7278-5116, Fan Xia, Ph.D. https://orcid.org/0000-0002-4974-9851, Hugo J. Bellen, D.V.M., Ph.D. https://orcid.org/0000-0001-5992-5989 [email protected], Pengfei Liu, Ph.D. https://orcid.org/0000-0002-4177-709X [email protected], and Zhandong Liu, Ph.D. https://orcid.org/0000-0002-7608-0831 [email protected]Author Info & Affiliations
Published April 25, 2024
NEJM AI 2024;1(5)
DOI: 10.1056/AIoa2300009

Abstract

Background

Diagnosing genetic disorders requires extensive manual curation and interpretation of candidate variants, a labor-intensive task even for trained geneticists. Although artificial intelligence (AI) shows promise in aiding these diagnoses, existing AI tools have only achieved moderate success for primary diagnosis.

Methods

AI-MARRVEL (AIM) uses a random-forest machine-learning classifier trained on over 3.5 million variants from thousands of diagnosed cases. AIM additionally incorporates expert-engineered features into training to recapitulate the intricate decision-making processes in molecular diagnosis. The online version of AIM is available at https://ai.marrvel.org. To evaluate AIM, we benchmarked it with diagnosed patients from three independent cohorts.

Results

AIM improved the rate of accurate genetic diagnosis, doubling the number of solved cases as compared with benchmarked methods, across three distinct real-world cohorts. To better identify diagnosable cases from the unsolved pools accumulated over time, we designed a confidence metric on which AIM achieved a precision rate of 98% and identified 57% of diagnosable cases out of a collection of 871 cases. Furthermore, AIM’s performance improved after being fine-tuned for targeted settings including recessive disorders and trio analysis. Finally, AIM demonstrated potential for novel disease gene discovery by correctly predicting two newly reported disease genes from the Undiagnosed Diseases Network.

Conclusions

AIM achieved superior accuracy compared with existing methods for genetic diagnosis. We anticipate that this tool may aid in primary diagnosis, reanalysis of unsolved cases, and the discovery of novel disease genes. (Funded by the NIH Common Fund and others.)

Understand the impact of AI on your practice. Get the tools to evaluate the risks.

Monthly & Annual Options

Notes

A data sharing statement provided by the authors is available with the full text of this article.
Supported by the NIH Common Fund, through the Office of Strategic Coordination/Office of the National Institutes of Health (NIH) Director under award number(s) NIH P50HD103555 to Z.L.; NIH/NHGRI R01 HG011795 to H.J.B., M.F.W., and Z.L.; NIH/NINDS U54 NS093793 and U54 NS093793-07S2 to H.J.B., S.Y., and M.F.W.; NIH/NHGRI R35HG011311 to P.L.; U01HG007709, U01HG007942, U01HG007943, U01HG007530, U01HG007672, U01HG010218, U01HG007708, U01HG007703, U01HG007674, and U54NS093793 to UDN. This research is also supported by the Chan Zuckerberg Initiative (grant no. 2023-332162 to Z.L., S.Y., and H.J.B.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. Z.L., D.M., C.L., and L.W. are partially supported by the Chao Endowment and the Huffington Foundation. Z.L., D.M., C.L., L.W., and H.J.B. are also supported by Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital.
Disclosure forms provided by the authors are available with the full text of this article.

Supplementary Material

Supplementary Appendix (aioa2300009_appendix.pdf)
Disclosure Forms (aioa2300009_disclosures.pdf)
Data Sharing Statement (aioa2300009_data-sharing.pdf)

Information & Authors

Information

Published In

History

Submitted: June 28, 2023
Revised: November 26, 2023
Accepted: January 29, 2024
Published online: April 25, 2024
Published in issue: April 25, 2024

Topics

Authors

Affiliations

Department of Pediatrics, Baylor College of Medicine, Houston
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Graduate School of Biomedical Sciences, Program in Quantitative and Computational Biosciences, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Graduate School of Biomedical Sciences, Program in Quantitative and Computational Biosciences, Baylor College of Medicine, Houston
Department of Pediatrics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Department of Data Science and AI, Al Hussein Technical University, Amman, Jordan
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Sasidhar Pasupuleti, M.S. https://orcid.org/0000-0003-0310-5812
Department of Pediatrics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Department of Pediatrics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Department of Pediatrics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Jill A. Rosenfeld, Ph.D. https://orcid.org/0000-0001-5664-7987
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Baylor Genetics, Houston
Lindsay C. Burrage, M.D., Ph.D. https://orcid.org/0000-0002-5108-8861
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Michael F. Wangler, M.D. https://orcid.org/0000-0001-5245-5910
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Shinya Yamamoto, D.V.M., Ph.D.* https://orcid.org/0000-0003-2172-8036 on behalf of for Undiagnosed Diseases Network,*
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Baylor Genetics, Houston
Baylor Genetics, Houston
Baylor Genetics, Houston
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Baylor Genetics, Houston
Brendan Lee, M.D., Ph.D. https://orcid.org/0000-0001-8573-4211
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Human Genome Sequencing Center, Baylor College of Medicine, Houston
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Baylor Genetics, Houston
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston
Department of Neuroscience, Baylor College of Medicine, Houston
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston
Baylor Genetics, Houston
Department of Pediatrics, Baylor College of Medicine, Houston
Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston

Notes

Dr. H. Bellen can be contacted at [email protected]; Dr. P. Liu can be contacted at [email protected]; and Dr. Z. Liu can be contacted at [email protected] or at Jan and Dan Duncan Neurological Research Institute, 1250 Moursund St., Houston, TX 77030-3411.
*
A complete list of members of the Undiagnosed Diseases Network is provided in the Supplementary Appendix, available at ai.nejm.org.
Drs. Mao, C. Liu, and Wang contributed equally to this work.

Metrics & Citations

Metrics

Altmetrics

Citations

Export citation

Select the format you want to export the citation of this publication.

View Options

View options

PDF

View PDF

Full Text

View Full Text

Media

Figures

Other

Tables

Share

Share

CONTENT LINK

Share

Evaluating clinical AI: Be empowered with the evidence you need.

Monthly & Annual Options

Subscribe Now