2
$\begingroup$

Is there a methodology for detecting whether a genome has been tampered with by a biologist? For instance, say there has been an ebola outbreak such as in "The Andromeda Strain". Is there a way a researcher can tell by just looking at the ebola genetic data and determine it is a genetically engineered strain?

Some ideas I have:

  1. organism is abnormally close to unrelated organisms, i.e. some of the genetic code from the other organisms has been inserted into the first

  2. subsequences are very different from the rest of the genome. If there is a consistent rate of mutation applied to the entire genome, then a subsequence that is abnormally noise free could be inserted.

  3. combination of 1 and 2: finding clear instances of unrelated organism genomes inserted into the genome being examined.

Additionally, are there any publicly accessible reference genomes that are known to be genetically engineered, that I could use as examples to test such ideas?

UPDATE: the wikipedia article on detecting genetically modified organism is not quite what I am looking for, since it claims current methods cannot detect unknown genetic modification.

https://en.wikipedia.org/wiki/Detection_of_genetically_modified_organisms

Currently, it is highly unlikely that the presence of unexpected or even unknown GMOs will be detected, since either the DNA sequence of the transgene or its product, the protein, must be known for detection.

$\endgroup$
5
  • 1
    $\begingroup$ Researcher have generated transgenic organisms for over a century. So yes there are many "modified genomes" available. You could also take any strain of interest, modify a few nucleotides, add some indels, or even an entire gene and analyze that. For step 3 for example, you could take humanized mouse or fly models used in disease biology... $\endgroup$
    – Mack123456
    Feb 14 '20 at 19:46
  • $\begingroup$ @MichaelG. since you reference conspiracy theories, probably in regard to the coronavirus outbreak: for the record, I do not think the Wuhan coronavirus is genetically engineered. But, the conspiracy theories made me wonder if there is a reliable way to detect whether genetically engineering has occurred in genetic data. $\endgroup$
    – yters
    Feb 14 '20 at 20:03
  • $\begingroup$ For the record, there is no guaranteed way to do this, it'd be completely impossible. We genetically engineer things all the time by just changing a few bases to things that already naturally occur in the species (but not at the same time in the wild) to speed up animal/crop breeding, as a simple example. $\endgroup$
    – Devon Ryan
    Feb 14 '20 at 20:10
  • $\begingroup$ @DevonRyan can we at least guarantee true positives with high reliability? It seems to be the case, e.g. if we see a massive segment of DNA from a totally different species spliced into the DNA sequence we are examining. I'm asking if there is a good general methodology for at least identifying true positives reliably, even if we cannot identify true negatives reliably. $\endgroup$
    – yters
    Feb 14 '20 at 21:22
  • $\begingroup$ Sure, transgenes will make things obvious, but they're not generally present when we genetically modify something. $\endgroup$
    – Devon Ryan
    Feb 14 '20 at 21:26
2
$\begingroup$

The definition of a GMO for example UK GMO regulatory bodies (HSE) is strictly the artificial introduction of foreign genetic material using targeted genetic engineering, such as splicing, genome integration etc... Attenuation for example by serial passage the basis of the yellow fever vaccine, does not count as a GMO.

The simple answer is look at phylogenetic trees between the GMO strain against the total distribution of isolates. If large regions have been spliced the 'silent mutation rate' looks very different,.because the GMO parent is from a very different environment to the recipient. The switch in replication rate affects the numbers of 'silent mutation', which is fast in RNA viruses.

If the GMO is a point mutation, ie limited amino acid change then I've actually done this with dengue virus and it is fairly easy because the novel mutation make the GMO 'stick out'.of the phylogeny against all natural mutation patterns. At this point the outlier is tracked back to Genbank where the relavant publication defines the engineered virus.

If you are inferring a conspiracy theory that is a different process and a different discussion, particularly if it is Ebolavirus which is a Category 4 pathogen. The investigation is again within the phylogeny, but in this case the answer is always the isolate is a natural zoonose from a sylvatic (wild) transmission cycle which has spilled over into humans.

$\endgroup$
5
  • $\begingroup$ Is there any way to do this if all you have is the particular strain and no database to compare it against? $\endgroup$
    – yters
    Feb 16 '20 at 19:07
  • $\begingroup$ No there is no internal signature. You need to compare the putative engineered (virus) against all other isolates, wild-type viruses. $\endgroup$
    – M__
    Feb 16 '20 at 19:35
  • $\begingroup$ Hmm, but we can detect plagiarism in this way. For example, let's say I have a long string of text and Lorem ipsum dolor sit amet, consectetur adipiscing elit. :) It seems like we could determine the characteristics of much of the sequence and then detect deviations from these characteristics, at least if the deviations are large enough. $\endgroup$
    – yters
    Feb 17 '20 at 19:25
  • $\begingroup$ Plagiarism is detected by reference to prior publication, same idea with DNA/RNA $\endgroup$
    – M__
    Feb 17 '20 at 20:12
  • $\begingroup$ I'm referring to stylometry: en.wikipedia.org/wiki/Plagiarism_detection#Stylometry. If something like this is not done in bioinformatics, seems like a good research avenue. $\endgroup$
    – yters
    Feb 18 '20 at 0:09
0
$\begingroup$

If you need to engineer a trait, then it has not evolved, so it probably does not confer a selectional advantage to the organism.

So a fingerprint of an engineered organism is that the genetic difference between it and its relatives does not have a selectional explanation.

$\endgroup$
1
  • $\begingroup$ Most of the things in nature don't have a selectional explanation. $\endgroup$
    – Kamil S Jaron
    Dec 1 '20 at 10:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.