<p>Various types of genetic constructs are widely used as diagnostic, prophylactic, and therapeutic tools for human diseases. They are also the workhorse in biotech and pharmaceutical industry for production of therapeutic antibodies and proteins. Since the majority of the genetic constructs encode protein products, it is therefore of tremendous value to human health and the society that we could find a way to fine-tune and optimize genetic constructs and hence protein expression for achieving maximal potency or long-lasting effects in therapeutics or for obtaining highest yields in pharmaceutical protein production. However, for protein-coding genes to be expressed in a heterologous host, the coding sequences need to be optimized by using synonymous codons to achieve reasonable levels of expression, if at all. Since codon optimization is done in a protein-by-protein basis with respect to specific host organisms, tissue/cell types, even health conditions, and there is no set of standard rules to follow, this process is still very unpredictable and time-consuming.</p><p>This thesis presents the development of a feasible platform for solving the problem of optimizing regular and long DNA constructs for academic or industrial purposes through the development of a novel cloning method for complex gene libraries, and based on the library expression system constructed in such manner, a platform for high-throughput screening of codon-optimized and error-corrected proteins, and a novel protocol for screening long gene constructs which could be extremely difficult to achieve by using regular screening methods. This multi-step platform has the potential for studying the natural systems: how codon bias correlates to protein expression efficiency, for generating improved pharmaceutical proteins and enhanced DNA vaccines and for constructing improved genome libraries.</p> / Dissertation