Additional file 14: of Evaluation of off-target and on-target scoring algorithms and integration into the guide RNA selection tool CRISPOR

Scatterplots of predicted versus obtained cleavage efficiency for all studies shown in Figs. 4 and 5 and Additional file 13: Figure S5 for all scoring models, including 3′ GC content-related heuristics. One row per dataset, one column per prediction score. The previously published datasets shown are, in order: the training data from the Wang, Doench 2014 (human and mouse) and Chari (293 T and K562) studies, Wang 2015, Doench 2016 (AZD and PLX treatment), Hart 2016 averaged over all time points (Rpe1 and the two Hct116 replicates), the data from studies by Moreno-Mateos, Varshney, Gagnon, Liu, Ren, Housden, Farboud, and Gandhi. For details and references on these datasets, see “Methods” or Additional file 10: Table S5. The last six rows represent datasets from this study: “Concordet” is a cell culture dataset quantified with T7 endonuclease and on gels, the datasets “Schönig” and “Eschstruth” ranked guides 1–3 by efficiency, “Shkumatava” by number of mutated sequencing clones obtained from zebrafish embryos, “Teboul” the percentage of mutant mouse embryos. For details on these datasets, see the “Methods” section and also Additional file 15: Table S7 for the Teboul dataset. (TSV 8773 kb)