Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Phenotypic signatures in clinical data enable systematic identification of patients for genetic testing


Around 5% of the population is affected by a rare genetic disease, yet most endure years of uncertainty before receiving a genetic test. A common feature of genetic diseases is the presence of multiple rare phenotypes that often span organ systems. Here, we use diagnostic billing information from longitudinal clinical data in the electronic health records (EHRs) of 2,286 patients who received a chromosomal microarray test, and 9,144 matched controls, to build a model to predict who should receive a genetic test. The model achieved high prediction accuracies in a held-out test sample (area under the receiver operating characteristic curve (AUROC), 0.97; area under the precision–recall curve (AUPRC), 0.92), in an independent hospital system (AUROC, 0.95; AUPRC, 0.62), and in an independent set of 172,265 patients in which cases were broadly defined as having an interaction with a genetics provider (AUROC, 0.9; AUPRC, 0.63). Patients carrying a putative pathogenic copy number variant were also accurately identified by the model. Compared with current approaches for genetic test determination, our model could identify more patients for testing while also increasing the proportion of those tested who have a genetic disease. We demonstrate that phenotypic patterns representative of a wide range of genetic diseases can be captured from EHRs to systematize decision-making for genetic testing, with the potential to speed up diagnosis, improve care and reduce costs.

This is a preview of subscription content, access via your institution

Relevant articles

Open Access articles citing this article.

Access options

Rent or buy this article

Prices vary by article type



Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Predictive performance of the model in a held-out CMA test set and a general hospital population.
Fig. 2: Identification of patients with CNV syndromes and interpretability.
Fig. 3: Proportion of patients with a putative pathogenic CNV identified by the model.
Fig. 4: Prediction performance across diverse genetic diseases.
Fig. 5: Clinical time period preceding the genetic test.

Data availability

Summary level data on frequency and importance of phecodes in the model are presented in Supplementary Table 3. Summary data on clinical and genetic information are provided throughout the paper. All requests for raw (for example CNV and phenotype) data and materials are reviewed by Vanderbilt University Medical Center to determine whether the request is subject to any intellectual property or confidentiality obligations. For example, patient-related data not included in the paper may be subject to patient confidentiality. Any such data and materials that can be shared will be released via a material transfer agreement. ClinGen data were downloaded from UCSC Genome Browser June 2019 ( DECIPHER CNV syndromes were extracted from

Code availability

All code used to construct and run the model is provided at


  1. Nguengang Wakap, S. et al. Estimating cumulative point prevalence of rare diseases: analysis of the Orphanet database. Eur. J. Hum. Genet. 28, 165–173 (2020).

    Article  Google Scholar 

  2. Ferreira, C. R. The burden of rare diseases. Am. J. Med. Genet. A 179, 885–892 (2019).

    Article  Google Scholar 

  3. Rosenthal, E. T., Biesecker, L. G. & Biesecker, B. B. Parental attitudes toward a diagnosis in children with unidentified multiple congenital anomaly syndromes. Am. J. Med. Genet. 103, 106–114 (2001).

    Article  CAS  Google Scholar 

  4. About Rare Diseases (Orphanet, accessed June 2020);

  5. About Rare Diseases (EURORDIS Rare Diseases Europe, accessed June 2020);

  6. Suther, S. & Kiros, G.-E. Barriers to the use of genetic testing: a study of racial and ethnic disparities. Genet. Med. 11, 655–662 (2009).

    Article  Google Scholar 

  7. Noll, A. et al. Barriers to Lynch syndrome testing and preoperative result availability in early-onset colorectal cancer: a national physician survey study. Clin. Transl. Gastroenterol. 9, 185 (2018).

    Article  Google Scholar 

  8. Moreno-de-Luca, D. et al. Clinical genetic testing in autism spectrum disorder in a large community-based population sample. JAMA Psychiatry 77, 979–981 (2020).

    Article  Google Scholar 

  9. OMIM: Online Mendelian Inheritance in Man (Johns Hopkins University, accessed June 2020);

  10. McKusick, V. A. Mendelian Inheritance in Man and its online version, OMIM. Am. J. Hum. Genet. 80, 588–604 (2007).

    Article  CAS  Google Scholar 

  11. Cooper, D. N., Krawczak, M., Polychronakos, C., Tyler-Smith, C. & Kehrer-Sawatzki, H. Where genotype is not predictive of phenotype: towards an understanding of the molecular basis of reduced penetrance in human inherited disease. Hum. Genet. 132, 1077–1130 (2013).

    Article  CAS  Google Scholar 

  12. Girirajan, S. et al. Phenotypic heterogeneity of genomic disorders and rare copy-number variants. N. Engl. J. Med. 367, 1321–1331 (2012).

    Article  CAS  Google Scholar 

  13. Posey, J. E. et al. Resolution of disease phenotypes resulting from multilocus genomic variation. N. Engl. J. Med. 376, 21–31 (2017).

    Article  CAS  Google Scholar 

  14. Goldstein, B. A., Navar, A. M., Pencina, M. J. & Ioannidis, J. P. A. Opportunities and challenges in developing risk prediction models with electronic health records data: a systematic review. J. Am. Med. Inform. Assoc. 24, 198–208 (2017).

    Article  Google Scholar 

  15. Bastarache, L. et al. Improving the phenotype risk score as a scalable approach to identifying patients with Mendelian disease. J. Am. Med. Inform. Assoc. 26, 1437–1447 (2019).

    Article  Google Scholar 

  16. Bastarache, L. et al. Phenotype risk scores identify patients with unrecognized Mendelian disease patterns. Science 359, 1233–1239 (2018).

    Article  CAS  Google Scholar 

  17. Denny, J. C. et al. Systematic comparison of phenome-wide association study of electronic medical record data and genome-wide association study data. Nat. Biotechnol. 31, 1102–1111 (2013).

    Article  CAS  Google Scholar 

  18. Firth, H. V. et al. DECIPHER: Database of Chromosomal Imbalance and Phenotype in Humans Using Ensembl Resources. Am. J. Hum. Genet. 84, 524–533 (2009).

    Article  CAS  Google Scholar 

  19. Corbett-Davies, S. & Goel, S. The measure and mismeasure of fairness: a critical review of fair machine learning. Preprint at arXiv (2018).

  20. Lundberg, S. M. et al. From local explanations to global understanding with explainable AI for trees. Nat. Mach. Intell. 2, 56–67 (2020).

    Article  Google Scholar 

  21. McInnes, L., Healy, J. & Melville, J. UMAP: uniform manifold approximation and projection for Dimension reduction. Preprint at arXiv (2018).

  22. Brokamp, C. et al. Material community deprivation and hospital utilization during the first year of life: an urban population-based cohort study. Ann. Epidemiol. 30, 37–43 (2019).

    Article  Google Scholar 

  23. Breiman, L. Random forests. Mach. Learn. 45, 5–32 (2001).

    Article  Google Scholar 

  24. Wang, K. et al. PennCNV: an integrated hidden Markov model designed for high-resolution copy number variation detection in whole-genome SNP genotyping data. Genome Res. 17, 1665–1674 (2007).

    Article  CAS  Google Scholar 

  25. Diskin, S. J. et al. Adjustment of genomic waves in signal intensities from whole-genome SNP genotyping platforms. Nucleic Acids Res. 36, e126 (2008).

    Article  Google Scholar 

  26. Amemiya, H. M., Kundaje, A. & Boyle, A. P. The ENCODE blacklist: identification of problematic regions of the genome. Sci. Rep. 9, 9354 (2019).

    Article  Google Scholar 

Download references


This work was supported by R01MH111776 (to D.M.R.), R01MH113362 (to N.J.C. and D.M.R.), R01LM010685 (to L.B.) and U01HG009068 (to N.J.C.). This study makes use of data generated by the DECIPHER community. A full list of centers that contributed to the generation of the data is available from and via email from Funding for the project was provided by Wellcome. The dataset(s) used for the analyses described were obtained from Vanderbilt University Medical Center’s BioVU, which is supported by institutional funding, private agencies and federal grants. These include the NIH-funded Shared Instrumentation Grant S10RR025141, and CTSA grants UL1TR002243, UL1TR000445 and UL1RR024975. Genomic data are also supported by investigator-led projects that include U01HG004798, R01NS032830, RC2GM092618, P50GM115305, U01HG006378, U19HL065962 and R01HD074711; and additional funding sources listed at

Author information

Authors and Affiliations



D.M.R. and T.J.M. designed and conceived the study. L.B. and T.J.M. extracted data from the EHRs for training and validation. L.H. generated the CNV data. T.J.M., D.M.R. and J.M. designed and implemented the prediction model. T.J.M. and D.M.R. performed the analyses. V.M.C. and R.H.P. performed external validation at MGB. D.M.R., T.J.M., L.B. and N.J.C. interpreted the results. T.J.M. and D.M.R. drafted the paper. All authors read the paper, provided feedback and approved the submission.

Corresponding author

Correspondence to Douglas M. Ruderfer.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer reviewer information Nature Medicine thanks Marc Williams and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Michael Basson was the primary editor on this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 PheWAS of CMA cases versus matched controls.

PheWAS Manhattan plot showing significance of associations from logistic regressions of each of 1,620 phecodes and whether an individual received a CMA vs controls. Triangle points represent direction of effect and points are colored by phecode category. For clarity, only phecodes with uncorrected p-values below 5 × 10−150 are labeled.

Extended Data Fig. 2 Age of patients at date of CMA testing differs by syndrome.

Age of patients at the time of their CMA report grouped into the most common syndromic region by combining diagnosis and genomic coordinates of reported abnormal variant. Independent patient numbers within each category: 15q11.2 syndromes (32), 16p11.2 syndromes (14), 1q21.1 syndromes (9), CMT/HNPP (18), DiGeorge/22q11.2 Duplication syndrome (31), Down Syndrome (7), Turner/Klinefelter (14), Williams syndrome (9).

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Morley, T.J., Han, L., Castro, V.M. et al. Phenotypic signatures in clinical data enable systematic identification of patients for genetic testing. Nat Med 27, 1097–1104 (2021).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

This article is cited by


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing