Nested pool testing strategy for the reliable identification of individuals infected with SARS-CoV-2

Abstract

AbstractThe progress of the SARS-CoV-2 pandemic requires the design of cost-effective testing programs at large scale. To this end, pooling multiple samples can provide a solution. Defining a cost-effective strategy requires the establishment of an efficient deconvolution and re-testing procedure that eventually allows the identifcation of the carrier. Based on Dorfman’s algorithm, we developed an adaptive nested strategy for which we have, for a given prevalence, simple analytic expressions of the optimal number of samples in the starting pool, of the number of partitioning steps (stages) in the optimal path, of the pool sizes in each of these stages and of the expected average number of tests needed to identify the infected individuals. In this paper we analyze the strategy in detail focusing on its practical implementation when there are restrictions that prevent the use of the optimum. More specifically, we analyze how to proceed when the infection prevalence is poorly known a priori or when the optimal requires starting with pool sizes that are too large for the reliable detection of an infected sample. The sensitivity of the RT-qPCR assay, the gold standard RNA detection method, is a major concern in the case of SARS-CoV-2: it is estimated that half of the infected individuals give false negative results. Recently, droplet digital PCR (ddPCR) was shown to be 10 − 100 times more sensitive than RT-qPCR, making this technology suitable for pool testing. ddPCR has the added value of providing the direct quantification of the RNA content at the end of the test. In the paper we show how this feature can be used for verification purposes. The analyses and strategies presented here should be useful to those considering the adoption of a pooling approach for RNA detection, particularly, for the identification of individuals infected with SARS-CoV-2.Author summaryThe progress of the SARS-CoV-2 pandemic requires the design of cost-effective testing programs at large scale. Running tests on pooled samples can provide a solution if the tests sensitivity is high enough. In the case of SARS-CoV-2, the current gold standard test, RT-qPCR, has shown some limitations that only allow the use of pools with relatively few samples. In this regard, Droplet digital PCR (ddPCR) has been shown to be 10 − 100 times more sensitive than RT-qPCR, making it suitable for test pooling. In this paper we describe a nested pool testing method in which the properties that make it optimal are simple analytic functions of the infection prevalence. We discuss how to proceed in practical implementations of the strategy, particularly when there are constraints that prevent the use of the optimal. We also show how its nested nature can be combined with the direct RNA quantification that the ddPCR test provides to identify the presence of unviable samples in the pools and for self-consistency tests. The studies of this paper should be useful for those considering the adoption of test pooling for RNA detection.