Together with the global scientific effort to establish and improve laboratories’ technical ability to perform precise tests on pooled samples, the dynamic strategies proposed in the doctoral students’ paper promise to significantly increase the efficiency of the testing process, which is crucial to the struggle against the global pandemic.
High volumes of Polymerase Chain Reaction (PCR) tests are a central tool in the global efforts to contain the novel coronavirus pandemic (COVID-19). Many countries are facing a shortage of labor, materials and equipment necessary to execute these tests. These shortages are not expected to be overcome in the very near future, but may in fact worsen, as emerging from the global crisis may demand mass-testing of general population, such as at schools, universities, and places of work.
One method that has recently been proposed to confront the burden of enormous numbers of tests is the method of sample pooling. The concept underlying this method is to combine samples from multiple patients into a single test in the laboratory. If this test gives a negative result, one can conclude that all the combined samples are negative. However, if this test gives a positive result, additional tests must be performed until the individual positive samples are identified. This method can significantly increase the number of tests that can be performed, and may prove critical in overcoming the unprecedented burden confronting laboratories across the world.
But how many samples should laboratories combine into a single batch? If they combine too many, it is overwhelmingly likely that each batch tested will be positive and they will have to perform a large number of additional tests in order to identify the positive individuals. On the other hand, if they combine only a small number of samples, the method will hardly reduce the total number of tests.
Four advanced doctoral students in the program for computational neuroscience at ELSC published a paper at medRxiv that offers computational tools to solve this problem. The authors, Itamar Landau, Haran Shani-Narkiss, Nadav Yayon, and Omri Gilday, used principles from information theory in order to produce practical protocols to enable laboratories to dynamically select the optimal batch size for pooling samples under changing circumstances, and to calculate the expected reduction in the total number of tests. The frequency of positive tests varies significantly from country to country and under differing testing conditions. In consideration of these differing testing conditions, the paper provides laboratories with a straightforward table that associates the optimal batch size with the changing conditions.