I am currently thinking about the difficulty of detecting and identifying the genome-edited (crispr/cas mediated) crops. If the changes are only few nucleotides I think they will be indistinguishable from natural variants (mutations) and the genetic modification technique cannot be 'identified'.
Does anyone has ideas of available (or new) detection methodologies? Can SNP-detection methods be applied (for 'knowledge-based' detection)? Is NGS or WGS applicable as 'non-targeted' method for detection/identification of genome-edited crops in the future?
Background: The US Department of Agriculture (USDA) decided in April this year that the first CRISPR-edited crops (a maize and a mushroom) can be cultivated and sold without regulatory oversight (Nat Biotechnol. 34(6):582). In the EU and Germany the discussion is ongoing how the regulatory system for the new plant breeding technologies will look like. The current EU system requires the availability of detection methods for the approval of GMOs under the EU legislation.
You're right, it can't formally be "proven" that a small deletion is CRISPR-caused and not natural. However, viable guide-RNA binding sites are not super-frequent in the genome. If you use common prediction tools, or experimentally validated libraries, you'll get maybe a handful of target sites per gene. If you just so happen to find a few base pairs deletion exactly in the place of a predicted gRNA site, in a gene that somebody would plausibly want to modify for an agricultural purpose, then that should probably count as a "smoking gun" (strongly suggestive evidence that CRISPR has been used). A natural mutation would seem quite unlikely to fall exactly upon a predicted CRISPR target site.
You're right, it can't formally be "proven" that a small deletion is CRISPR-caused and not natural. However, viable guide-RNA binding sites are not super-frequent in the genome. If you use common prediction tools, or experimentally validated libraries, you'll get maybe a handful of target sites per gene. If you just so happen to find a few base pairs deletion exactly in the place of a predicted gRNA site, in a gene that somebody would plausibly want to modify for an agricultural purpose, then that should probably count as a "smoking gun" (strongly suggestive evidence that CRISPR has been used). A natural mutation would seem quite unlikely to fall exactly upon a predicted CRISPR target site.
Dear John
I want to know more information about these gRNA sites. ¿Would you suggest some reference?
Thank you,
Sofía.
One easy and frequently used example is the Zhang lab prediction software http://crispr.mit.edu/. They have published a lot of papers explaining how it works. This is a reference explaining this type of software: https://www.ncbi.nlm.nih.gov/pubmed/24956386
- Badges
- Science method