Comments on: At BIO 2010!

By: Ewan R

Ewan R — Fri, 07 May 2010 20:48:32 +0000

One issue with this, and it’s a perception issue rather than necessarily a science one – looking at TxP or MxP you can essentially guarantee statistically significant differences. The Kogel paper for instance (if it’s the one I’m thinking of) shows that there are differences but that varietal differences are at least as big as, if not bigger than, differences seen in the transgenic.

Now, any scientist (other than a handful I guess… waiting for a Seralini response on this, or a UCS white paper…) would look at this and conclude that there aren’t significant differences – whereas I can totally see the anti-GM side leaping all over it and screaming about 2% of transcripts/metabolites being significantly different therefore this isn’t substantial equivalence.

To confound things, you could conceivably have a plant with very small changes in TxP and MxP being non-equivalent (lets say we engineer in THC or something) and with massive TxP and MxP differences being substantially equivalent in terms of nutrition etc (admittedly it’s somewhat harder to see how something massively deviant would be substantially equivalent)

You set your panel too wide you find stuff. You find stuff once, you give ammunition to the other side. Even if it never repeats. The toxicity studies by Monsanto reanalysed by Seralini show just this – results which show something to be non-dangerous still fuel magical thinking due to failure to consult a statistician (or at least one who knows his arse from his elbow)

By: Anastasia Bodnar

Anastasia Bodnar — Wed, 05 May 2010 17:31:07 +0000

Welcome I have a few more where those came from, I’ll post ‘em later on.
I think the targets will change as the method becomes more sophisticated. It might make sense to target specific groups of compounds (or specific groups of genes) at first, then as we get better at identifying all those peaks, start looking at more groups. But overall, I think looking at sig differences in the transgenic variety compared to a wide panel will at least give us a start on where to look – to see if those differences are biologically significant as well as statistically significant.

By: pdiff

pdiff — Wed, 05 May 2010 15:19:26 +0000

Anastasia,
I’m glad you posted these references. I do a bit of work in metabolomics (the word is fine, I just hate saying it!) and these will be interesting reads. One of the problems I would see with this application is identifying the signals you get. Many times, in the species I’ve seen (brassicas), the mix from the mass spec is quite complex. You have to know what you are looking for. The mass spec is picking up a lot of intermediate precursor compounds, etc. Do you envision going after specific classes of compounds, e.g. a class of proteins or other by products?

By: Anastasia Bodnar

Anastasia Bodnar — Wed, 05 May 2010 14:56:29 +0000

Frank, I have a Q I’d like to you to ask of a few different people when you get a chance. When I was at BIGMAP, I talked to some biotech regulators about using metabolomics or transcriptomics to determine substantial equivalence. Basically, the idea is that you could use these tools to compare a transgenic event, its nontransgenic isoline, and a diverse panel from the same species – then see if anything in the transgenic one is significantly different from levels in the others. Some examples so far have been barley (Kogel, 2010) and potatoes (Lehesranta, 2005). I personally think this has potential for 1) testing transgenic varieties to find out if there really are any differences, 2) actually learning what the differences are, if any, and 3) boosting consumer confidence because it would allow us to definitively say “there are no unintended differences”. I’d so far as to say this could replace a lot of the tests that are currently being done and could streamline the regulatory process. But the people I talked to said 1) the tests aren’t valuable enough to warrant the price, 2) expression from plant to plant of the same line or even the same plant at different times differs too much so we can’t depend on differences between a GM plant and its isoline, 3) toxicology tests were more valuable. I think they’re being short sighted, and I’d like to know what others say about it.

PS: Brandon says I should’t talk like this, he doesn’t like the word metabolomics :p