CROP: A feature-independent context-aware method for CRISPR-Cas9 frameshift prediction

Abstract

MotivationThe CRISPR-Cas9 complex has revolutionized genome-editing technologies. By designing a 20 nt-long guide RNA, a Cas9 nuclease can be guided to cleave almost any genomic target site (followed by NGG). The cleavage induces double-stranded DNA breaks, which are then repaired by cellular pathways. Accurate CRISPR-Cas9 repair-outcome prediction is essential for designing guide RNAs with desired genomic effects, such as gene knockout. A central challenge is quantifying the rate of frameshifts, i.e. repair-outcomes that lead to a change in the local length that is not a multiplicity of 3. Previous methods for frameshift-rate prediction were trained on only few experimental or cellular contexts, mostly rely on manually defined microhomology features, and are limited by sparse features and class labels. ResultsWe developed CROP, the first feature-independent context-aware repair-outcome prediction method. By aggregating specific repair outcomes as {Delta}length classes, CROP overcomes class sparsity. We designed CROP to work with variable input sequence lengths and output classes in order to utilize multiple datasets simultaneously. We benchmarked CROP against state-of-the-art repair-outcome prediction methods over 18 datasets, which we curated and standardized from various studies. Across all datasets, CROP outperformed all competing methods in frameshift-rate prediction. We performed cross-experiment and cross-cellular frameshift predictions to investigate the generalizability of repair mechanisms. Finally, we show that CROP learned microhomology principles from raw sequences without explicit feature engineering, establishing the first end-to-end architecture for CRISPR-Cas9 repair-outcome prediction which learns from multiple datasets. Availability and implementationCROP is available at https://github.com/OrensteinLab/CROP.