We demonstrate a highly parallel strategy to analyze the impact of single nucleotide mutations on protein function. Using our method, it is possible to screen a population and quickly identify a subset of functionally interesting mutants. Our method utilizes a combination of yeast functional complementation, growth competition of mutant pools, and polymerase colonies. A defined mutant human glucose-6-phosphate-dehydrogenase library was constructed which contains all possible single nucleotide missense mutations in the eight-residue glucose-6-phosphate binding peptide of the enzyme. Mutant human enzymes were expressed in a zwf1 (gene encoding yeast homologue) deletion strain of Saccharomyces cerevisiae. Growth rates of the 54 mutant strains arising from this library were measured in parallel in conditions selective for active hG6PD. Several residues were identified which tolerated no mutations (Asp200, His201 and Lys205) and two (Ile199 and Leu203) tolerated several substitutions. Arg198, Tyr202, and Gly204 tolerated only 1-2 specific substitutions. Generalizing from the positions of tolerated and non-tolerated amino acid substitutions, hypotheses were generated about the functional role of specific residues, which could, potentially, be tested using higher resolution/lower throughput methods.