Probabilistic Modelling of Base Editing Efficiency and Bystander Outcome Composition
- 1University of Oxford
Abstract
Base editors install single nucleotide substitutions without a double strand break or donor template, but editing at multiple susceptible positions can generate bystander alleles. Predicting the frequency of an intended outcome therefore requires both editing efficiency and the composition of edited products. Here we present crispAIBE, a probabilistic framework that uses separately trained transformer convolutional networks to predict Dirichlet distributions for editing efficiency and conditional allele composition. Split-conformal density-level regions are calibrated on data reserved from model selection, with composition thresholds fitted by outcome-set size. On held-out data, crispAIBE achieves Pearson correlations of 0.669 for efficiency and 0.896 for composition on measured allele sets, with empirical coverages of 0.902 and 0.906 at a nominal level of 0.90. On the published CRISPRon-BE benchmark, with test protospacers excluded from training, crispAIBE ranks first in seven of eight comparisons under the two-column Pearson statistic and first or second in seven of eight under the rank statistic. Within predicted-efficiency quintiles, the largest- and smallest-region deciles differ 1.9-fold in mean absolute error on average, compared with 1.7-fold for a boundary-distance ranking. When evaluated on U2OS, which was excluded from training, larger predictive regions were associated with larger absolute prediction errors. Local calibration with 100 U2OS protospacers yields median empirical coverage of 0.918 at the nominal 0.90 level. The crispAIBE repository is https://github.com/furkanozdenn/crispAIBE.
Related
- Analysis of SpCas9 on- and off-target effects in high efficiency multiplex editing in Arabidopsis
- Promoter editing of OsGTL1 improves instantaneous water use efficiency via targeted stomatal reduction in rice
- High efficiency gene editing and phenotype generation in avian embryos using lipid nanoparticles as a vector delivery platform
- Genetics of energy allocation to growth in Angus cattle: trait derivation, model comparison, and relationships with feed efficiency and body composition traits
- Macrophage-mediated bystander T cell activation in varicose veins
BibTeX
@misc{ozden2026probabilistic,
title = {{Probabilistic Modelling of Base Editing Efficiency and Bystander Outcome Composition}},
author = {F. Ozden and P. Minary},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.09.757916},
url = {https://www.biorxiv.org/content/10.64898/2026.10.09.757916v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Probabilistic Modelling of Base Editing Efficiency and Bystander Outcome Composition
AU - F. Ozden
AU - P. Minary
PY - 2026
DA - 2026/10/09
PB - bioRxiv
DO - 10.64898/2026.10.09.757916
UR - https://www.biorxiv.org/content/10.64898/2026.10.09.757916v1
AB - Base editors install single nucleotide substitutions without a double strand break or donor template, but editing at multiple susceptible positions can generate bystander alleles. Predicting the frequency of an intended outcome therefore requires both editing efficiency and the composition of edited products. Here we present crispAIBE, a probabilistic framework that uses separately trained transformer convolutional networks to predict Dirichlet distributions for editing efficiency and conditional allele composition. Split-conformal density-level regions are calibrated on data reserved from model selection, with composition thresholds fitted by outcome-set size. On held-out data, crispAIBE achieves Pearson correlations of 0.669 for efficiency and 0.896 for composition on measured allele sets, with empirical coverages of 0.902 and 0.906 at a nominal level of 0.90. On the published CRISPRon-BE benchmark, with test protospacers excluded from training, crispAIBE ranks first in seven of eight comparisons under the two-column Pearson statistic and first or second in seven of eight under the rank statistic. Within predicted-efficiency quintiles, the largest- and smallest-region deciles differ 1.9-fold in mean absolute error on average, compared with 1.7-fold for a boundary-distance ranking. When evaluated on U2OS, which was excluded from training, larger predictive regions were associated with larger absolute prediction errors. Local calibration with 100 U2OS protospacers yields median empirical coverage of 0.918 at the nominal 0.90 level. The crispAIBE repository is https://github.com/furkanozdenn/crispAIBE.
ER -