Design of a nucleic acid library is an essential foundation for an in vitro selection. Increasing the starting diversity of a pool will increase the sampling of sequence space but other strategies must be considered to access rare secondary structures including varying random region length, modularity of the secondary structure and introducing complexity. Efforts to understand these characteristics are reviewed.

In vitro evolution of peptides and proteins has been enabled by display technologies such as phage and mRNA display, but each has limitations. To bypass these, a new system that utilizes peptide-DNA conjugations was designed by employing in vitro compartmentalization, which colocalized genotype and phenotype allowing selections. This system could permit a 1014 starting diversity. To optimize DNA and RNA stability in the coupled transcription/translation reaction and the compartmentalized reaction, DNA constructs were designed to code for reporter peptides and altered to increase mRNA stability and translation yields. Pool construction was then optimized and the flanking fixed sequences as well as four bioorthogonal moieties were installed. A mock selection was optimized using an electrophoretic shift mobility assay (EMSA) leading to the same result in the positive and negative controls, which suspended the project.

In vitro selection of a ssDNA pool was performed in the presence of fluorescein mono-β-D-galactopyranoside (FMG) and fluorescein di(β-D-galactopyranoside (FDG), substrates for a β-galactosidase. Twenty-one rounds of selection were performed with increasing stringency including a decrease of substrate to 2 µM, a decrease of incubation to 10s and addition of 7 M urea and incubation at 95 °C after the selection reaction. Fluorescence increases were observed when FDG and FMG were incubated with ssDNA from rounds 19 and 21 as well as with individual DNA clones from round 21. Incubation of individual clones with boronate affinity gel did not result in detectable binding, indicating that galactose had not been transferred from the substrate to the DNA clone. Reselection of the DNA pool was performed by competitively eluting bound ssDNA sequences with 0.1 M ribose but it did not result in enrichment of the DNA pool for active sequences over the six additional rounds of selection.