Non-random method of gene shuffling

The present invention concerns the non-random assembling of DNA molecules in a DNA construct and methods of using such constructs, including the production of nucleic acid libraries. The non-random gene shuffling is preferably accomplished by the following steps. First, optionally, the amino acid sequences of proteins encoded by related gene families of interest are aligned and inspected for regions of conserved amino acid residues. These conserved regions, preferably of at least 4 (e.g. about 4 to 10) consecutive conserved amino acid residues are candidate regions for the subsequent design of PCR primers to amplify the variable or less conserved regions in between them, followed by non-random reassembly to create a recombinant nucleic acid genetic library of gene family variants.