About the database

This database contains CRISPR targets for the mm10 genome. While some targets have been manually designed, most have been obtained from the automated whole-genome detection that we have introduced in our 2016 article.

If you use this database, please make sure to cite:
G. A. Sunagawa, K. Sumiyama, M. Ukai-Tadenuma, D. Perrin, H. Fujishima, H. Ukai, O. Nishimura, S. Shi, R. Ohno, R. Narumi, Y. Shimizu, D. Tone, K. L. Ode, S. Kuraku and H. R. Ueda (2016). "Mammalian reverse genetics without crossing reveals Nr3a as a short-sleeper gene". Cell Reports 14:1-16.

The article details the entire selection method. Candidates are first extracted from a gene's exons, based on their sequence, and then go through successive filtering steps. Candidates are eliminated if they appear multiple times in the genome, if their AT content is below 45%, if they contain TTTT, or if they are too close to the reverse primer for PCR amplification in the gRNA-template construction. They are also discarded if the corresponding gRNA has a poor secondary structure, or if the sequence has a high off-target risk. Candidates that pass all these steps are considered suitable targets, and are stored in a database.
This can be summarised by the following figure, reproduced from our artcile.

[Panel A, Figure 7]

The database provides targets for 21,559 genes. For 19,536 genes, we have identified at least three targets. This means the triple-targeting method outlined in the article can be applied to 81.2% of the mm10 genome using this database.
Moreover, 71.9% of the gene has more than six independent targets, which corresponds to more than two sets of triple-target CRISPR gRNAs.

[Panel B, Figure 7]

Copyright © RIKEN Quantitative Biology Center, 2015.