TIDE vs Surveyor

[regulatingge...@gmail.com]

Hi,
After transfecting cells with Cas9 and the gRNAs. I picked colonies and checked the indels using both Surveyor and TIDE. With TIDE, the best colony has a 90% efficiency. With Surveyor, I did not get any cleavage for any of the colonies. I ran the control reaction and I had a clear cleavage band on the gel. Do I trust TIDE and proceed with the clone that had 90% efficiency for functional assays? or is surveyor more reliable?
Thanks.

[Victor Dillard]

Hi, Surveyor and endonuclease mismatch assays have been shown to be less reliable than quantitative methods like TIDE or NGS-based methods. Guides with high activity (>90% NHEJ) as calculated with TIDE have been shown to also not show much or any activity on T7 gels.

- Victor.

[kevin....@synthego.com]

Agreed - the cleavage assays grossly underestimate the amount of actual cutting efficiency and NHEJ repair compared to TIDE analysis. 

- Kevin

[Vanessa Y]

Hi, 

I'm running into the same issue of which one is more reliable. Also are you having trouble accessing their web tool? I have not been able to open it the last couple of hours. 

[regulatingge...@gmail.com]

Hi,
Thanks for your post. How can I tell if my cells are homozygous or heterozygous from TIDE?
Also, if they are heterozygous, does the protein level go down by half or does it vary from gene to gene? Should I run a western blot to assess protein levels?
Finally, does using NGS provide more conclusive information about the location of the indel in comparison to using Sanger sequencing with TIDE?
Thanks.

[Mark]

Can you please give more details on this experiment, like what do you mean by picking colonies? Are these ES cells grown on a feeder layer or something like that?

[Mike]

Hi,

Vanessa: the site seems to be live where I am (UK).

Regulating: this is a bit of a simplistic view. A monoclonal cell line targeted by CRISPR is likely to have any of three outcomes: Homozygous wild-type (which will show close to 100% '0' on the indel spectrum report in TIDE), heterozygous wild-type and indel (which will show two peaks both around 50% on the indel spectrum, one at '0' and one indel of a certain size), or compound heterozygous with two indels both of 50% and no wild-type. It is unlikely that both alleles will have the same indel, but TIDE would be able to suggest this, as you would have close to 100% on the indel spectrum at a site other than '0'. Of course NGS is the gold standard, as it can show you not just what is there and at what proportion but the exact change that occurred. For example a -2 indel in TIDE might actually be two different indels, one that deleted two bases from the left of the cut site, and one that deleted one base from the left and one base from the right of the cut site. TIDE is blind to these, as well as indels where there is a both a deletion and insertion. But you get a lot of information quickly, easily and inexpensively with TIDE.

Cheers,
Mike

[regulatingge...@gmail.com]

Hi Mike,
Thanks for the clarification. I have 2 more questions:
1. Is there a protocol for NGS for checking CRISPR. Do I just isolate the genomic DNA and hand it over to our NGS facility?
2. On TIDE, I have 60% at the -2 indel position and 30% at the -3 indel position. Does this mean that I have indels at the 2 alleles? I copied the table from TIDE below. I also do not understand why the highest 2 percentages (indel -2 and -3) have a p-value of 0?

Quantification Indel Frequencies

overall efficiency = 91.8 %

	percentage	pvalue
-10	0.0	1
-9	0.0	1
-8	0.0	1
-7	0.0	1
-6	0.0	1
-5	0.0	1
-4	0.0	1
-3	30.8	0
-2	60.5	0
-1	0.0	1
0	2.6	3.9e-07
+1	0.0	1
+2	0.0	1
+3	0.0	1
+4	0.5	0.34
+5	0.0	1
+6	0.0	0.99
+7	0.0	0.95
+8	0.0	1
+9	0.0	1
+10	0.0	1

Thanks.

[regulatingge...@gmail.com]

Hi Mark,
These are standard cell lines. I diluted the transfected cells at about 10 cells per mL and plated them in 96 well plates

[kevin....@synthego.com]

Also good to keep in mind that sometimes, TIDE will not show an Indel, but at some targets, it's possible after NHEJ for the repair to look scarless, even after a CRISPR cut.  Just something to keep in mind.  It can be hard to see anything higher than 80% Indels.  If you are doing HDR, or if you have some other kind of readout (e.g., GFP, cell sorting) you may see higher editing results.

[Mark]

Thanks reg. I only ask because if the population is indeed monoclonal I'd expect a maximum of only two types of indels, roughly 1:1 or 100% if they're the same class (rare). Mike already stated this above. Of course this is only assuming there are two copies of that gene in your cell. Personally I think western blot or some functional assay is the best way to test for complete knockout. For NGS I guess it will depend on what your facility offers in terms of sample prep, but generally you will at least have to PCR the targeted site and then add the NGS adapters, or barcodes if you want to pool samples and whatnot.

[Andy Grierson]

Probably worth checking karyotype/cytogenetic data for cell lines and your target gene before starting. Most of the standard lab cell lines, like HeLa, carry many additional and rearranged chromosomes. This doesn't stop them being useful tools for cell biology, but it will certainly complicate CRISPR-mediated editing.

HeLa zygosity is detailed in Figure 1 of

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3737162/

[Joe Miano]

Totally agree; we spent a year trying to target a supposed diploid cell line only to find, upon karyotyping, the line is tetraploid!

Strongly recommend karyotyping cell lines, even those that are indicated as aneuploid since we saw an instance where there were 6 copies of a single chromosome!