The abstract is read at the publisher; the summary is JClub's.
A convolutional neural network, 2CaP-5, predicts LbCas12a crRNA editing efficiency in wheat protoplasts with moderate to strong performance across ~1,000 crRNAs.
- Why it matters: Existing Cas12a RNA activity models lack plant-specific training data and show limited predictive power in plant systems, hindering efficient crRNA selection.
- What they did: The study used a high-throughput arrayed wheat protoplast transfection assay to measure on- and off-target editing for ~1,000 unique LbCas12a crRNAs, then tested 40 deep learning architectures to train 2CaP-5.
- The result: 2CaP-5 did not generalize to Agrobacterium-mediated transformation, but it could be re-trained with as few as 500 crRNAs and, using ~2,000 wheat subgenome off-targets, confirmed that single mismatches near the PAM reduce indel efficiency and three mismatches can largely prevent off-target activity.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.