Single-cell RNA-seq data are challenging because of the sparseness of the read counts, the tiny expression of many relevant genes, and the variability in the efficiency of RNA extraction for different cells. We consider a simple probabilistic model for read counts, based on a negative binomial distribution for each gene, modified by a cell-dependent coefficient interpreted as an extraction efficiency. We provide two alternative fast methods to estimate the model parameters, together with the probability that a cell results in zero read counts for a gene. This allows to measure genes co-expression and differential expression in a novel way.