Directed evolution represents a versatile tool to tailor enzyme properties to needs in industrial applications and to understand structure-function relationships. Genetic diversity is commonly generated using error-prone PCR. Exploration of sequence space by random mutagenesis strongly favors transitions when enzyme-based mutagenesis methods are employed (Wong, T. S., Zhurina, D., Schwaneberg, U., Comb. Chem. High Throughput Screen. 2006, 9, 271-288). The genetic code has been organized in a manner that limits chemical diversity when a single transition mutation occurs in a codon (Wong, T. S., Roccatano, D., Schwaneberg, U., Biocatal. Biotransformation 2006, in press). Are transitions more beneficial than transversions for adapting biocatalysts to non-natural process conditions? In a statistical analysis performed with the Mutagenesis Assistant Program (MAP), we compared the consequences of transition and transversion bias on amino acid substitution patterns of the P450 BM-3 heme domain. For the analysis, we used a recently introduced benchmarking system consisting of a protein structure indicator, an amino acid diversity indicator with a codon diversity coefficient, and a chemical diversity indicator. A detailed analysis for the P450 BM-3 heme domain showed that an ideal transversion bias generates more diverse amino acid substitution patterns with a significantly different chemical composition than an ideal transition bias. Emphasis is given on the theoretical analysis with a brief discussion on potential implication of transition and transversion bias in directed evolution experiments.