Published Paper
Download Model
GitHub Repository (legacy)
The links marked legacy point to the original published source code. For the model as this web server runs it, use the Download Model link above.
Deep learning-based base editor and Cas9 variant guide RNA design and efficiency prediction tool, published in Kim et al. Nature Biotechnology (2024).
DeepBE and DeepCas9variants design guide RNAs and predict the efficiency of base editors and Cas9 variants. The DeepBE and DeepCas9variants deep learning models were generated based on the measured activities of various base editors (ABE8e(V106W), ABE8.17m+V106W, YE1-BE4max, SsAPOBEC3B, CGBE1, miniCGBE1, and APOBEC-nCas9-Ung, fused to each of the Cas9 variants mentioned next) and Cas9 variants (SpCas9, SpCas9-NG, VRQR variant, SpCas9-NRRH, SpCas9-NRTH, SpCas9-NRCH, SpG, SpRY, and Sc++).
Inputs that specify an edit. An unedited/edited sequence pair, ClinVar ID and COSMIC ID are accepted here as ways of naming the edit you want: the target window is resolved from the variant, and each candidate is reported against the conversion it would have to make. No homology-directed repair (HDR) donor is designed on this workflow — a base editor deaminates within its own activity window and makes no double-strand break for a donor to repair from — so when the requested conversion cannot be reached from that window, the run names the editor, its window and the conversion asked for, instead of returning an empty design.
