Sequence Display and protein AI broaden compact SlugCas9 targeting in cultured cells
Published in Artificial Intelligence, Bioinformatics, Journals, Tools.
Published November 1, 2026.
A Nature Communications study published on October 30 applies Sequence Display and protein language models to expand the DNA sequences accessible to a compact CRISPR enzyme.
Cheng and colleagues mapped how individual and combined mutations alter the activity of SlugCas9 across four related PAM contexts: NNAA, NNAT, NNAC and NNAG. PAMs are short neighboring DNA sequences that constrain where Cas9 can act.
Combining experimental sequence–activity data with predictive models identified variants with improved and broader activity across these contexts. The variants also improved base editing at selected endogenous sites in cultured mammalian cells, including sites in APOC3 and PCSK9.
The advance concerns reprogramming targeting preferences, rather than removing every targeting constraint. Results at selected cell-culture sites do not establish genome-wide specificity, safe delivery to patients or therapeutic effectiveness. This application builds on earlier Sequence Display work and should not be read as the platform’s first introduction.
The paper discloses a patent application involving Sequence Display-enabled machine learning. The peer-reviewed accepted article remains subject to editorial changes.
Source: Cheng et al., Nature Communications, October 30, 2026.