RNA interference (RNAi) is a phenomenon whereby expression of an individual gene is specifically silenced by the introduction of a double-stranded RNA (dsRNA) whose sequence is homologous to that of the gene in question. The generation of a small interfering RNA (siRNA) expression library directed against the entire human genome is a project that requires solutions to many difficult technical problems. We present here some strategies for solving some of these problems, including the development of genetically stable and highly active siRNA expression vectors, a procedure for selection of favorable target sites, and an efficient and inexpensive procedure for constructing an siRNA expression library.