Summary of techniques frequently applied in directed evolution.

	Technique	Purpose	Advantages	Disadvantages	Application examples
Mutagenesis	Error-prone PCR and error-prone RCA	Insertion of point mutations across whole sequence	• Easy to perform	• Reduced sampling of mutagenesis space	Subtilisin E12
			• Does not require prior knowledge about key positions	• Mutagenesis bias	Glycolyl-CoA carboxylase124
	RAISE	Insertion of random short insertions and deletions	• Enables random indels across sequence	• Indels limited to few nucleotides
				• Frameshifts introduced	β-Lactamase17
	TRINS	Insertion of random tandem repeats	• Mimics duplications that occur in natural evolution	• Frameshifts introduced	β-Lactamase18
	Mini-mu based techniques	Random insertion or deletion of one or multiple codons	• Conservation of reading frame		Arylesterase18
			• Modification of transposon enables customization of mutations	• Additional steps of DNA manipulation required	GFP125
	Mutator strains	In vivo random mutagenesis	• Simple system	• Biased and uncontrolled mutagenesis spectrum	Vitamin K epoxide reductase126
				• Mutagenesis not restricted to target	Cells resistant to DMF127
	Orthogonal systems based on DNA Pol I, pGLK1/2, Ty1, T7RNAP and CRISPR	In vivo random mutagenesis	• Mutagenesis restricted to target sequence	• Mutation frequency relatively low	β-Lactamase28
			• Could be coupled to in vivo selection	• Limitations on size of target sequence	Dihydrofolate reductase128
					Orotidine-5′-phosphate decarboxylase31
	DNA shuffling	Random sequence recombination	• Recombination advantages	• High homology between parental sequences required	Thymidine kinase129
					Non-canonical esterase130
	StEP	Random sequence recombination	• Recombination advantages	• High homology between parental sequences required	Photoswitchable fluorescent protein131
			• Easy to perform		Synthetic antibodies39
	RACHITT	Random sequence recombination	• Increased crossover frequency		Monooxygenase40
			• Parental sequences removed from library	• High homology between parental sequences required	DNA polymerase132
Mutagenesis	ITCHY and SCRATCHY	Random recombination of any two sequences	• No homology between sequences required	• Gene length and reading frame not preserved	Alcohol dehydrogenase41
				• Recombination at sites not structurally related	Deoxyribonucleoside kinase133
				• Single crossover per variant (solved in SCRATCHY)
	SHIPREC	Random recombination of any two sequences	• No homology between sequences required	• Single crossover per variant
			• Crossovers at structurally-related sites	• Reading frame not preserved	Cytochrome P45042
	SISDC	Recombination of a few sequences at specific sites	• No homology between sequences required	• Limited number of potential crossover points
			• Crossover points can be chosen	• Additional steps of DNA manipulation required	β-Lactamase44
	MORPHING	In vivo generation of recombination libraries	• Can be coupled to in vivo selection techniques	• Crossovers only occur at overlapping regions	Peroxidase45
					Aryl alcohol oxidase134
	Site-saturation mutagenesis	Focused mutagenesis of specific positions	• In-depth exploration of mutagenesis at chosen positions	• Only a few positions mutated	Widely applied to enzyme evolution56,57
			• Possibility to incorporate previous information for efficient mutagenesis	• Libraries can easily become very large
			• Iterative cycles and smart libraries can reduce library sizes
	StLois	Sequential extension of loops	• Insertions performed at sites less likely to result in non-functional variants
			• Reasonable library sizes due to sequential extension of loops	• Limited number of insertions in each extension cycle	Cumene dioxygenase64
Identification of variants	Colorimetric/fluorimetric analysis of colonies/cultures	Screening of variants	• Fast and easy to perform	• Limited to biomolecules exhibiting appropriate spectral properties	Fluorescent proteins57,58
	Plate-based automated enzymatic assays	Screening of variants	• Automation has increased throughput	• Throughput remains limited compared to other methods, especially if substrate or product do not have characteristic spectral or fluorescent properties
			• Surrogate substrates expand scope	• Results with surrogate substrates do not always replicate with original ones	Lipase135
			• Coupling to GC/HPLC enables analysis of enantiomers		Laccase136
	FACS-based methods	Screening of variants	• High throughput	• Evolved property must be linked to a change in fluorescence	Sortase81
			• Product entrapment expands application scope		Cre recombinase82
			• Similar techniques can be applied with in vitro compartmentalization		β-galactosidase83
	MS-based methods	Screening of variants	• High throughput	• Less widely-available equipment required	Fatty acid synthase77
			• Does not rely on specific properties of substrates	• For MALDI-based methods, requirement of immobilization on matrix	Cytochrome P41178
					Cyclodipeptide synthase79
	Display techniques	Selection of variants	• High throughput	• Limited to selection of biomolecules with specific binding properties	Antibodies87
					Fbs1 glycan-binding protein90
					RNA-binding peptides137
					Random sequence ATP-binding proteins138
	QUEST	Selection of variants	• High throughput	• Limited scope due to substrate/ligand constraints	Scytalone dehydratase101
					Arabinose isomerase139
	Cofactor regeneration coupling	Selection of variants	• High throughput		Alcohol dehydrogenase109
			• Applicable to wide range of small molecule biocatalysts and properties	• An indirect link to NAD-related activities must be established	Imine reductase109
					Nitrorreductase109
					Isopropanol pathway109
	In vitro compartmentalized self-replication	Selection of variants	• High throughput	• Limited to activities that can be linked to replication or transcription of its coding sequence	DNA polymerase104
			• Bypasses library transformation