Deep learning prediction of bacterial genetic engineering using a hybrid CNN-transformer approach
Proceedings of the Canadian Conference on Artificial Intelligence · 20.05.2025
DOI: 10.21428/594757db.87d49597 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the Canadian Conference on Artificial Intelligence, online erschienen am 20.05.2025. Verfasst von Tyler Browne.
Abstract (Originalsprache)
The Cas9 enzyme alongside a single guide RNA molecule provides a modular tool for genetic engineering and, more recently, has shown its effectiveness as a species-specific anti-microbial. The Cas9 enzyme is targeted to cut DNA by a programmable guide sequence. Though many target sites exist, activity varies depending on the site selected. For practical usage, the ability to accurately predict on-target activity is desired. We propose a hybrid CNN-transformer model, built on a yet unused dataset of target site sequences and activity scores. At present, model performance exceeds all prior models across all tested datasets, functioning well across different bacterial species. Through further tuning and the implementation of an ensemble methodology, we aim to provide a robust cross-species bacterial sgRNA-Cas9 prediction model that provides genetic engineers with confidence that Cas9 will perform as desired.
Zitieren
APA
Browne, T. (2025). Deep learning prediction of bacterial genetic engineering using a hybrid CNN-transformer approach. Proceedings of the Canadian Conference on Artificial Intelligence. https://doi.org/10.21428/594757db.87d49597
Vancouver
Browne T. Deep learning prediction of bacterial genetic engineering using a hybrid CNN-transformer approach. Proceedings of the Canadian Conference on Artificial Intelligence. 2025. doi:10.21428/594757db.87d49597
BibTeX
@article{Browne2025,
author = {Browne, Tyler},
title = {Deep learning prediction of bacterial genetic engineering using a hybrid CNN-transformer approach},
journal = {Proceedings of the Canadian Conference on Artificial Intelligence},
year = {2025},
doi = {10.21428/594757db.87d49597}
}
EUVIMED