Additional file 1 of Improved prediction of bacterial CRISPRi guide efficiency from depletion screens through mixed-effect machine learning and data integration

Additional file 1: Figure S1. Illustration of the genomic and sequence features used. Figure S2. Comparison of guide depletion across datasets. Figure S3. Spearman correlation of 10-fold cross-validation of models trained with one or mixed datasets. Figure S4. Data integration for retrained Pasteur and deep learning models. Figure S5. Interaction between distance features and whether targeting gene is the first gene in operon. Figure S6. Independent low-throughput validation of model performance. Figure S7. Additional figures related to the saturating screen of purine biosynthesis genes. Figure S8. Model performance of deep learning approaches. Supplementary Note: Deep learning approaches do not improve prediction performance.