Abstract:
|
Array-based group testing algorithms, where individual specimens are tested in overlapping pools initially, are widely used in diseases screening, drug discovery and genetics. It has proven to be a cost-efficient method when screening a large number of individuals for a binary characteristic of interest. Motivated by the increasingly use of the multiplex assays that can detect multiple infections simultaneously, we generalize previous statistical work in array-based group testing algorithms with single infection to allow for the presence of multiple infections. We first derive the operating characteristics of the proposed algorithms, including the expected number of tests and classification accuracy probabilities, and then compare them with previously studied hierarchical algorithms which account for multiple infections. We also identify the optimal group testing algorithms that minimize the expected number of tests. The performance of our methods is illustrated via simulation and the chlamydia and gonorrhea data collected in Nebraska as part of the Infertility Prevention Project.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.