Defining treshold for gene expression in single cell RNA-seq

Question

[asente]

What would be a standard way to define a raw count treshold above which it would be safe to assume a gene is expressed in the particular cell based on single-cell RNAseq counts? Many thanks.