In the 'post-genomic' era, cDNA and genomic sequences are now available that encode huge numbers of proteins. Assigning functions to these proteins is a daunting task. Cell biologists have traditionally approached this problem by disrupting protein function with dominant-negative or structural mutants. Here, we describe several alternative approaches whereby cells or cell lines lacking particular gene products can be generated from genomic sequences for use in functional studies. These include gene targeting in mouse, human and chicken DT40 cells, and recent advances in double-stranded RNA-mediated interference (RNAi).