Between the 3rd of October and 6th of November  2018 the text for this clause read as follows

  • The DARPA program is easily weaponized (this is nothing to do with the fact that DARPA is a military agency)

This in part appeared to form the basis of Washington Post Editorial on the 4th Nov 2018 where it was asserted by the Editorial Board that “The skeptics say on their website that they worry the DARPA program could be “easily weaponized.” That seems a stretch.” .  

We appreciate that the word ‘easily’ may have been better chosen to reflect the text used in of the Science article.  

Whether or not a chromosomal editing system (e.g., CRISPR) is ultimately used to achieve DARPA’s stipulated aims, easy simplifications (and not elaborations) of the described work program could be used to generate a new class of biological weapon.


In our view, the program is primarily a bad idea because obvious simplifications of the work plan with already-existing technol- ogy (10, 11) can generate predictable and fast-acting weapons, along with their means of delivery, capable of threatening virtually any crop species (see the figure). 

In providing simplified statements for non-specialists this may have led us to inadequately qualify this clause. We have now done so.

We would however note  that we had also clearly provide  the scientific basis (including a key citation)  for part of our opinion on this website - sadly this was neither rebutted or even discussed by the authors of the Washington Post Opinion piece.

Why do we assert that it is almost always easier to develop a HEGAA biological weapon system than an agricultural one ? 

  • To conceive  of a routine use in agriculture for HEGAAs  it will be necessary to build in multiple mechanisms to control the spatial and taxonomic spread of the viruses.  This will be complicated.
  • When developing a biological weapon, most if not all ‘safeguards’ or ‘independent kill switches’ can be left out.  Usefully the targeting of gene editing to your enemies crops is assured by the DNA sequence specificity of the 2 guide RNAs used by CRISPR (see Figure in Science paper, it is notable that the specificity of the guide sequences is one of the few elements that the insect allies program relies on which is in any way largely predictable )
  • While to kill or sterilise a plant can likely be achieved by disrupting a single suitable gene, more complex traits will probably require  inserting new genes or genes into plant chromosomes (examples repeatedly presented include resistance to drought, frost, flooding, salinity, herbicides, and plant disease).  In most circumstances it is more than x1000 times more efficient to cause gene disruption than gene insertions (eg Mao, Z., Bozzella, M., Seluanov, A. & Gorbunova, V. Comparison of nonhomologous end joining and homologous recombination in human cells. DNA Repair (Amst.) 7, 1765–1771 (2008) ).

Have  genetically modified  viruses that have a capacity to edit chromosomes ever been developed? 

  • Yes. The following 2015 paper describes a genetically modified virus which when injected into the tails of laboratory mice can knock-out a specific gene in cells it infects.  In this laboratory confined experiment after a week >40% of the chromosomes in liver cells had been irreversibly disrupted.

Ran, F. A.; Cong, L.; Yan, W. X.; Scott, D. A.; Gootenberg, J. S.; Kriz, A. J.; Zetsche, B.; Shalem, O.; Wu, X.; Makarova, K. S.; Koonin, E. V.; Sharp, P. A.; Zhang, F. In vivo genome editing using Staphylococcus aureus Cas9. Nature 2015, 520, 186–191, doi:10.1038/nature14299.

Are there natural plant viruses that have a capacity to edit plant chromosomes? 

  • No.  Unlike animals plants do not have retroviruses that can alter their chromosomes.

Have  genetically modified  plant viruses that have a capacity to edit chromosomes in plant genomes ever been developed? 

  • NOT really. There are laboratory based protocols that use viruses to gene edit plant chromosomes but these currently require the infiltration and a bacterium to be effective (it is not the spraying of a virus onto plants). In a laboratory setting this dependence is partly a safety feature providing little motivation to remove it. It does seam reasonable that with sufficient effort it would be possible to generate viruses which carry all the requirements for a  gene editing system.  It is however conceivably possible to split parts into different released viruses (slide 15) .Note plant viruses can be selected to have broad host ranges. Viruses use in laboratory techniques for germl ine plant manipulation can in their wildtype state infect 100s of species across many Families , e.g. the papers cited in the Science article 
  •  10. Z. Ali et al., Efficient Virus-Mediated Genome Editing in Plants Using the CRISPR/Cas9 System. Mol. Plant. 8, 1288-1291 (2015) .uses TRV= 400 species of plants from 50 families are susceptible to infection (nematode vector)
  • 11. K. Musiychuk et al., A launch vector for the production of vaccine antigens in plants. Influenza Other Respir. Viruses. 1, 19-25 (2007). uses TMV= nine plant families, and at least 125 individual species (insect vector).

No analytics are collected and no cookies are used on this site.