Novel Pooling Strategies for Genetic Testing, with Application to Newborn Screening

• 成果类型: Article
• 署名作者: El Hajj, Hussein; Bish, Douglas R.; Bish, Ebru K.; Kay, Denise M.
• 署名单位: University of Waterloo; University of Alabama System; University of Alabama Tuscaloosa; Wadsworth Center
• 刊物名称: MANAGEMENT SCIENCE
• ISSN/ISSBN: 0025-1909
• DOI: 10.1287/mnsc.2021.4289
• 发表日期: 2022
• 页码: 7994-8014
• 关键词: genetic testing; Newborn Screening; resource allocation; Pooled testing; partition problem

摘要

Newborn screening (NBS) is a state-level initiative that detects life-threatening genetic disorders for which early treatment can substantially improve health outcomes. Cystic fibrosis (CF) is among the most prevalent disorders in NBS. CF can be caused by a large number of mutation variants to the CFTR gene. Most states use a multitest CF screening process that includes a genetic test (DNA). However, due to cost concerns, DNA is used only on a small subset of newborns (based on a low-cost biomarker test with low classification accuracy), and only for a small subset of CF-causing variants. To overcome the cost barriers of expanded genetic testing, we explore a novel approach, of multipanel pooled DNA testing. This approach leads not only to a novel optimization problem (variant selection for screening, variant partition into multipanels, and pool size determination for each panel), but also to novel CF NBS processes. We establish key structural properties of optimal multipanel pooled DNA designs; develop a methodology that generates a family of optimal designs at different costs; and characterize the conditions under which a 1-panel versus a multipanel design is optimal. This methodology can assist decision-makers to design a screening process, considering the cost versus accuracy trade-off. Our case study, based on published CF NBS data from the state of New York, indicates that the multipanel and pooling aspects of genetic testing work synergistically, and the proposed NBS processes have the potential to substantially improve both the efficiency and accuracy of current practices.