Xue Zhong1,2, Zhijun Yin3,4, Gengjie Jia5, Dan Zhou6,7, Qiang Wei7,8, Annika Faucon9, Patrick Evans6,7, Eric R Gamazon6,7,10,11, Bingshan Li7,8, Ran Tao7,12, Andrey Rzhetsky5,13,14, Lisa Bastarache3, Nancy J Cox15,16. 1. Division of Genetic Medicine, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA. xue.zhong@vanderbilt.edu. 2. Vanderbilt Genetics Institute, Nashville, TN, USA. xue.zhong@vanderbilt.edu. 3. Department of Biomedical Informatics, Vanderbilt University Medical Center, Nashville, TN, USA. 4. Department of Electrical Engineering and Computer Science, Vanderbilt University, Nashville, TN, USA. 5. Department of Medicine, Institute of Genomics and Systems Biology, University of Chicago, Chicago, IL, USA. 6. Division of Genetic Medicine, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA. 7. Vanderbilt Genetics Institute, Nashville, TN, USA. 8. Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, TN, USA. 9. Human Genetics Graduate Program, Vanderbilt University, Nashville, TN, USA. 10. 'Life Member' of Clare Hall, University of Cambridge, Cambridge, United Kingdom. 11. MRC Epidemiology Unit, University of Cambridge, Cambridge, United Kingdom. 12. Department of Biostatistics, Vanderbilt University Medical Center, Nashville, TN, USA. 13. Committee on Genomics, Genetics and Systems Biology, University of Chicago, Chicago, IL, USA. 14. Department of Human Genetics, University of Chicago, Chicago, IL, USA. 15. Division of Genetic Medicine, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA. nancy.j.cox@vanderbilt.edu. 16. Vanderbilt Genetics Institute, Nashville, TN, USA. nancy.j.cox@vanderbilt.edu.
Abstract
PURPOSE: The increasing use of electronic health records (EHRs) and biobanks offers unique opportunities to study Mendelian diseases. We described a novel approach to summarize clinical manifestations from patient EHRs into phenotypic evidence for cystic fibrosis (CF) with potential to alert unrecognized patients of the disease. METHODS: We estimated genetically predicted expression (GReX) of cystic fibrosis transmembrane conductance regulator (CFTR) and tested for association with clinical diagnoses in the Vanderbilt University biobank (N = 9142 persons of European descent with 71 cases of CF). The top associated EHR phenotypes were assessed in combination as a phenotype risk score (PheRS) for discriminating CF case status in an additional 2.8 million patients from Vanderbilt University Medical Center (VUMC) and 125,305 adult patients including 25,314 CF cases from MarketScan, an independent external cohort. RESULTS: GReX of CFTR was associated with EHR phenotypes consistent with CF. PheRS constructed using the EHR phenotypes and weights discovered by the genetic associations improved discriminative power for CF over the initially proposed PheRS in both VUMC and MarketScan. CONCLUSION: Our study demonstrates the power of EHRs for clinical description of CF and the benefits of using a genetics-informed weighing scheme in construction of a phenotype risk score. This research may find broad applications for phenomic studies of Mendelian disease genes.
PURPOSE: The increasing use of electronic health records (EHRs) and biobanks offers unique opportunities to study Mendelian diseases. We described a novel approach to summarize clinical manifestations from patient EHRs into phenotypic evidence for cystic fibrosis (CF) with potential to alert unrecognized patients of the disease. METHODS: We estimated genetically predicted expression (GReX) of cystic fibrosis transmembrane conductance regulator (CFTR) and tested for association with clinical diagnoses in the Vanderbilt University biobank (N = 9142 persons of European descent with 71 cases of CF). The top associated EHR phenotypes were assessed in combination as a phenotype risk score (PheRS) for discriminating CF case status in an additional 2.8 million patients from Vanderbilt University Medical Center (VUMC) and 125,305 adult patients including 25,314 CF cases from MarketScan, an independent external cohort. RESULTS: GReX of CFTR was associated with EHR phenotypes consistent with CF. PheRS constructed using the EHR phenotypes and weights discovered by the genetic associations improved discriminative power for CF over the initially proposed PheRS in both VUMC and MarketScan. CONCLUSION: Our study demonstrates the power of EHRs for clinical description of CF and the benefits of using a genetics-informed weighing scheme in construction of a phenotype risk score. This research may find broad applications for phenomic studies of Mendelian disease genes.
Authors: Adam Lewis; Lisa Bastarache; Anita Pandit; Daniel B Larach; Jing He; Anik Sinha; Nicholas J Douville; Michael Heung; Michael R Mathis; Jonathan D Mosley; Jonathan P Wanderer; Sachin Kheterpal; Matthew Zawistowski; Chad M Brummett; Edward D Siew; Cassianne Robinson-Cohen; Miklos D Kertai Journal: BMC Nephrol Date: 2022-10-21 Impact factor: 2.585
Authors: Miklos D Kertai; Jonathan D Mosley; Jing He; Abinaya Ramakrishnan; Mark J Abdelmalak; Yurim Hong; M Benjamin Shoemaker; Dan M Roden; Lisa Bastarache Journal: Circ Genom Precis Med Date: 2021-03-01