Comprehensive evaluation and prediction of base editing outcomes for near-PAMless adenine and cytosine base editors

This dataset provides high-throughput sequencing results evaluating the editing efficiencies and outcomes of near-PAMless adenine and cytosine base editors. Experiments were performed in HEK293T cells using a library of 45,747 sgRNA-target sequences, including 24,050 random sequences, 1,023 sequences for evaluating PAM preferences, 20,541 ClinVar variants, and 133 endogenous loci. Multiple near-PAMless base editor variants were analyzed to elucidate sequence determinants of editing outcomes. The dataset further enables training and validation of computational model EnsemBE to accurately predict editing results for guiding design and therapeutic use of near-PAMless base editors.