We developed a novel method for constructing nearly random peptide library. Genomic DNAs extracted from tissue or cells of large genome species were digested with frequent cutter to produce short DNA fragments. These short fragments can be considered nearly random. Nearly random peptide libraries can be constructed by cloning the short fragments into appropriate expression vectors and transformation into host cells. Genomic DNA from one species can be digested with different restriction enzymes and ligated to different reading frames to produce several different libraries. In this study, we digested tobacco genomic DNA with two enzymes and cloned into three different reading frames to make totally six nearly random peptide libraries.