In functional genomics, the use of expression libraries of DNA variants in combination with potent screening techniques is a powerful tool for gene discovery. They allow study of gene and protein function, generation of peptide variants with novel properties, as well as identification of functional short DNA and RNA motifs. In proteomics, generation of large expression libraries of protein variants with random substitutions ("directed evolution") and further screening for novel or improved functions has been commonly used for isolation of proteins with novel characteristics, for improving enzymes, for rapid isolation of antibodies, and for functional protein studies. Most commonly, peptide libraries are expressed and screened in prokaryotic systems. Such systems have the advantage of rapid and simple generation of clones expressing single variants, allow high diversity (up to 10(11)), and can be combined with phage- or cell-surface display technique (2). The main disadvantage of bacterial systems is the absence of posttranslational modifications and native folding of many mammalian proteins, leading to limited applications, particularly when enzyme-substrate-, protein-protein, or protein-RNA interactions are to be studied.