101. Assessment of Published Bovine Guide RNA Design Practices.

Abstract

CRISPR/Cas9 is the most widely applied gene-editing (GnEd) technology in livestock, with more than 200 published studies reporting the generation of live, gene-edited animals for agricultural applications. This system consists of a Cas9 nuclease that cleaves DNA and a 19-20 base pair guide RNA (gRNA) that directs the nuclease and binds to a complementary genomic sequence. Accordingly, gRNA design is a primary determinant of editing outcomes, impacting both targeting specificity and cleavage efficiency. In this study, a literature search was conducted to identify published bovine gRNAs used in gene-editing studies targeting agriculturally relevant traits. Studies were required to be peer-reviewed and published in English. When ≥10 gRNAs were reported in a study, only those selected for downstream experimental use were included; otherwise, all reported gRNAs were retained. All gRNAs were aligned to the Bos taurus reference genome (ARS-UCD2.0 or the version specified in the study) using BLAST. Two in silico tools, Cas-OFFinder and CHOPCHOP, were used to predict gRNA binding specificity in the bovine reference genome (ARS UCD2.0 and ARS-UCD1.2.108, respectively). Each gRNA was evaluated using the commonly recommended selection benchmark of a minimum of three mismatches (non-consensus base pairings) between the gRNA and DNA at the most comparable predicted non-target loci, with at least one mismatch located in the seed region (the 8-11 bp adjacent to the nuclease recognition motif). Overall, a list of 70 gRNAs was curated. The analysis revealed considerable variability in gRNA design practices, with over half (52%, n = 37/70) of the published gRNAs not meeting the recommended benchmark for selection. Results from both prediction tools indicated that, on average, the most similar predicted non-target loci differed from the respective gRNA sequence by 2.7 mismatches (range: 1-4 mismatches), with three mismatches being the most common. However, one gRNA, based on Cas-OFFinder results, was found to have a predicted non-target locus that differed by only a single mismatch. Additionally, CHOPCHOP provides a predicted efficiency score for each gRNA ranging from 0-100 with higher scores indicating more efficient gRNAs. The CHOPCHOP efficiency scores for the curated guides ranged from 0 to 75.2 with an average score of 49. These results revealed considerable variability in design practices relative to commonly recommended mismatch-based criteria, highlighting opportunities to further standardize gRNA design approaches and develop practical tools compatible with current workflows that improve consistency in future livestock gene-editing applications.