Chromatin context-dependent regulation and epigenetic manipulation of prime editing

Abstract

Prime editing is a powerful means of introducing precise changes to specific locations in mammalian genomes. However, the widely varying efficiency of prime editing across target sites of interest has limited its adoption in the context of both basic research and clinical settings. Here, we set out to exhaustively characterize the impact of the cis-chromatin environment on prime editing efficiency. Using a newly developed and highly sensitive method for mapping the genomic locations of a randomly integrated "sensor", we identify specific epigenetic features that strongly correlate with the highly variable efficiency of prime editing across different genomic locations. Next, to assess the interaction of trans-acting factors with the cis-chromatin environment, we develop and apply a pooled genetic screening approach with which the impact of knocking down various DNA repair factors on prime editing efficiency can be stratified by cis-chromatin context. Finally, we demonstrate that we can dramatically modulate the efficiency of prime editing through epigenome editing, i.e. altering chromatin state in a locus-specific manner in order to increase or decrease the efficiency of prime editing at a target site. Looking forward, we envision that the insights and tools described here will broaden the range of both basic research and therapeutic contexts in which prime editing is useful.

Predicting prime editing efficiency across diverse edit types and chromatin contexts with machine learning

“…We incorporated all 288,793 pegRNAs, spanning both the Training and Test set, into the training process of the PRIDICT2.0 model. To validate ePRIDICT, we utilized the dataset from Li et al 19 , which includes 4144 integration locations and the corresponding prime editing efficiencies. Feature values were extracted from all averaging windows (100, 1000, 2000, and 5000 bp up and downstream) across 455 ENCODE datasets, following the same methodology applied to our TRIP library.…”

“…For further validation of ePRIDICT, we also applied the model on a recently generated independent prime editing dataset 19 . Similar to our study, Li et al 19 performed prime editing on a single target site that was integrated into different genomic loci in K562 cells.…”

“…For further validation of ePRIDICT, we also applied the model on a recently generated independent prime editing dataset 19 . Similar to our study, Li et al 19 performed prime editing on a single target site that was integrated into different genomic loci in K562 cells. Despite introducing another edit type (insertion instead of transversion) and targeting a different sequence, ePRIDICT reached a correlation of R=0.53, r=0.51 ( ExtFig4j ).…”

“…( i ) Spearman and Pearson correlation of XGBoost model prediction to editing efficiencies in TRIP library for prime editing, adenine base editing (ABE8e), cytosine base editing (BE4max), and SpCas9 genome editing. ( j ) Validation of ePRIDICT on an independent dataset from Li et al 19 , where one sequence was integrated and edited (CTT insertion) at 4144 genomic locations. ( h , j ) Color gradient from dark purple to yellow indicates increasing point density, per Gaussian KDE.…”

High throughput evaluation of genetic variants with prime editing sensor libraries

“…1e ). In general, the prime editing sensor seems to slightly overestimate the editing activity at the endogenous locus, likely in part due to differences in locus chromatin accessibility 42 , but the ranking of pegRNA editing efficiencies is largely preserved, validating our sensor-based approach.…”