ABSTRACT
This chapter summarizes statistical methods and models serving
as useful guides when designing and generating high-quality
mutant libraries in directed evolution, the emphasis being on
saturation mutagenesis as the preferred gene mutagenesis method.
Especially, the degree of oversampling when assessing mutant
libraries needs to be considered for optimal experimentation, as
screening is the main bottleneck in the overall procedure compared
to genetic selection and display systems. Thus, efforts toward
practical applications should be focused on reducing the screening
effort as much as possible. Along these lines, the statistical basis
when choosing appropriate multiresidue randomization sites in
conjunction with reduced amino acid alphabets is illuminated. A
comparison between the traditional Patrick and Firth algorithm
and the recently proposed statistical approach by Nov is presented.
Finally, practical tips on how to perform iterative saturation
mutagenesis (ISM) in the quest to evolve enzyme variants with
increased activity, enhanced or reversed stereoselectivity, altered
regioselectivity, and higher thermostability are also given.