Statistical prediction of the number of cells surviving infection by an autointerfering virus using the Poisson distribution

A method for estimating the number of defective interfering virus particles in a virus sample is presented. It can be used whenever the interference results in the survival of the “interfered” cell. The analysis assumes only that the infectious virus and defective interfering particles are distributed randomly and independently to cells. Thus the proportion of cells receiving X = x virus and Y = y particles is the product of the two independent Poisson distribution terms. The two dimensional matrix (X values × Y values) that can be constructed encompasses all of the possible (cellular) outcomes of viral infection. By comparing the actual number of surviving cells with the number predicted by various models of interference, it is possible to determine whether defective interfering particles are dominant (completely or partially) to infectious virus, and to estimate their number in the virus sample. This is accomplished by determining the experimental survival curve (% survival vs. input infectious virus/cell) and then constructing theoretical curves to fit the data.