Microarray studies require substantial amounts of RNA to generate probes, especially when bacterial RNA is used for hybridization. On page 679, Talaat et al. describe a computer-based algorithm for predicting the minimal number of primers to specifically anneal to all genes in a given genome. They show that genome-directed primers (GDPs) are more sensitive and specific than random-primer probes when hybridized to an array of 960 Mycobacterium tuberculosis genes. The approach could be useful for genome-wide expression analysis and directed amplification of sequenced genomes.