Fig. 2: Regeneration of ribozymes after cleavage enables selection and an NGS-based assay that are correlated with in vivo activity.

a CleaveSeq and DRIVER use the same method to regenerate the 5′ prefix after cleavage. b The regeneration method selectively restores the 5′ cleaved portion of the ribozyme and replaces the prefix sequence with a new prefix (e.g., “W” prefix is replaced with “Z” prefix for cleaved RNA molecules). The process starts with co-transcriptional cleavage of a DNA template library. An oligonucleotide is then added and annealed to the 3′ end of the resulting RNA pool for reverse transcription. The oligonucleotide subsequently hybridizes to the nascent cDNA, forming a partially self-annealing double-stranded hairpin that brings together the ends of molecules derived from the cleaved RNA, enhancing self-ligation. The circularized ligation product is cut at two uracil locations by Uracil-Specific Excision Reagent (USER), releasing a linear DNA strand harboring the desired sequence with a new prefix sequence. Two distinct populations of DNA molecules result: those corresponding to RNA that did not cleave and those that correspond to cleaved RNA, the latter of which will have the new prefix. One population is selectively PCR-amplified with primers that extend the product with either the T7 promoter (for DRIVER) or NGS adapters (for CleaveSeq). c CleaveSeq measures the relative abundance of cleaved and uncleaved molecules to provide estimates of cleavage fractions and switching for each library sequence. d Representative comparison of cleavage fractions for two replicates independently carried through the CleaveSeq assay (N = 12,025, at least 100 reads/sequence in each analysis). Bottom panel shows the standard error of the ratio on the y-axis. Significant (two-sided test with Bonferroni correction: p < 0.1/N) outliers are shown as red dots. Red dashed lines delineate 2.5-fold change of cleavage. e Comparison of CleaveSeq assay and gel electrophoresis analysis. Twenty-one biosensors, the hammerhead ribozyme, and a non-cleaving control were analyzed via CleaveSeq and on a 10% PAGE gel (Supplementary Figure 5). f Comparison of in vitro cleavage fraction (CleaveSeq) and in vivo gene regulatory activity. Each point represents an individual biosensor sequence. Gene regulatory activity, measured as the ability of the biosensor to control GFP reporter expression in yeast in a previously reported FACS-Seq assay²⁷ (N = 16,699). Green line is data processed by a 835-point median filter. Source data is available in the Source Data file.