Hi,

 CRISPRSeek, an R package for finding suitable guide RNAs (sgRNAs) for CRISPR targeting, has a lot of output files with partially-undocumented column names. These things may be obvious to a CRISPR expert, but for me, some of the columns are still a mystery.

 If anyone else has observations about the ideal output of this program, I would love to have their contributions here!

I went through and collected the ones I could understand, but still had some questions:

1. Sometimes "efficiency" is used as a term, and sometimes "efficacy" is used. I assume one is a typo.

2. sgRNA efficiency: this seems to be the primary output number. It is normalized from 0 to 1, which represents __________ (?). There is a linked paper that describes this, but I did not understand the math.

From the developers: a HIGHER number (closer to 1.0) is better: The higher the gRNA efficience the better it can target the intended site (see http://www.ncbi.nlm.nih.gov/pubmed/25184501)

3. Free energy: This seems to always be a NEGATIVE number, on a different scale as above. I am not clear what this scale is, or if this number is incorporated in sgRNA efficiency?

From the developers: The lower (more negative) the free energy the more stable the structure it is. There are three columns relevant to the free energy output in the summary file. You can type ?foldgRNAs in a R session to get some sense on what they are.

•  mfe.sgRNA    (free energy; "probability to form secondary structure with gRNA"—should be HIGH)
•  mfe.diff (should be HIGH: equal to mfe.sgRNA – mfe.backbone)
•  mfe.backbone (should be LOW: "probability to form secondary structure within the backbone itself")

Summary from the developers: You would ideally want mfe.sgRNA to be high, i.e., low probability to form secondary structure with gRNA plus the backbone is low, mfe.backbone to be low, i.e.,high probability to form secondary structure within the backbone itself, and the difference mfe.diff (mfe.sgRNA - mfe.backbone) to be high (less negative).

Note: Low here means a more negative number, e.g., -10 is lower than -1.

There is also a review that is relevant to this here: http://link.springer.com/article/10.1007%2Fs11515-015-1366-y ("Overview of guide RNA design tools for CRISPR-Cas9 genome editing technology")

Alex,

Thanks for taking time to post our email communications! 

Efficacy and Efficiency are used interchangeably here, referring to the cleavage likelihood for a given gRNA on a intended target. It ranges from 0 to 1 and the higher the efficacy, the more effective the gRNA.

So what are the rules that govern on-target cleavage efficacy?

The Root Lab studied the rules by creating a pool of 1841 sgRNAs, tiling across all possible target sites of a panel of six endogenous mouse and three endogenous human genes and quantitatively assessed their ability to produce null alleles of their target gene by antibody staining and flow cytometry.

The data from 1,841 sgRNAs were used to construct a model to predict the efficacy using sequence features of the expanded gRNA. 72 features were found statistically significant to contribute to the gRNA efficacy including GC content, some single nucleotide and dinucleotide variants  at some positions across the length of the gRNA and flanking target sequence. Minimum Free Energy (MFE) is not included as a predictive feature for the current model. For detailed information about the model, please refer to the paper from the Root lab (Doench, et al Nbt Aug 21, 2014). 

Best regards,

Julie