A systematic review of gene editing approaches supporting sustainable animal agriculture through molecular understanding and responsible practice

Abstract

Programmable genome editing enables heritable genetic modifications for livestock production, disease resistance, and animal welfare. However, its translational maturity across species and traits has not been systematically benchmarked against breeding-program standards. To address this gap, we conducted a systematic review following PRISMA 2020 guidelines. The Web of Science Core Collection (SCI-Expanded) was searched for peer-reviewed articles published between 2010 and 2026 using terms spanning ZFN, TALEN, CRISPR–Cas systems, base editors, prime editors, livestock, animal breeding, and biosafety. A total of 1,326 records were identified; 526 underwent full-text assessment, and 133 met eligibility criteria for qualitative synthesis. A quantitative subset of 28 primary experimental studies reporting editing efficiency, mosaicism, or multigenerational transmission was subjected to structured cross-study comparison. CRISPR–Cas9 was used in 24 of the 28 quantitative studies; notably, no primary livestock study reported founder-level quantitative data for base or prime editing. Across embryo-stage knockout studies, the median on-target efficiency was 84% (range: 45–100%). While SCNT-derived founders were uniformly non-mosaic by clonal design, mosaicism in zygote-edited animals ranged from 20 to 100%. Off-target analysis was reported in 43% of studies, but only 11% employed whole-genome sequencing (WGS). Multigenerational validation beyond founders was documented in 25% of studies, with three reaching the F2 generation or equivalent. Species-level maturity was highest for porcine CD163 editing, which demonstrated multigenerational, production-relevant validation for PRRSV resistance. No primary study identified in this review reported replicated parent–offspring trio-based WGS in directly zygote-edited livestock under contemporary methodological standards. For somatic cell nuclear transfer (SCNT)-derived founders, variant analysis was typically performed relative to the unedited donor cell line, and genome-wide assessments in this context have reported low off-target variant burdens relative to donor baselines, although cross-study comparability remains limited by heterogeneity in sequencing depth, variant-calling pipelines, and reporting thresholds. Finally, unintended on-target structural alterations, including plasmid backbone integration, were documented in SCNT-derived cattle. A robust proof-of-concept has been achieved for monogenic trait editing in pigs and cattle, yet the evidence base for breeding-scale deployment remains limited. Critical gaps include the absence of trio-based off-target baselines, a lack of cross-platform structural variant comparisons, and the unvalidated integration of computational tools within livestock editing pipelines. Responsible translation requires molecular validation depth proportional to dissemination intent as well as multi-generational, multi-environment performance data prior to commercialization.