Plants would be more reliable, though, since the plant roots actually penetrate the ground and the plant flowers (which are colored differently in the presence of mines) stick up where humans can see them.

Also, conceivably local fauna will convergently evolve with the plants and their DNA will “learn” to avoid the ones that show mines.

I suppose another angle is – if people are being forced by starvation to farm land that may be mined (or forced by the possibility of landmines to starve to death, I guess), then any measure that increases the likelihood of detecting mines is worthwhile.

http://www.cosmosmagazine.com/news/2151/genetically-modified-plants-detect-landmines

The objections people are voicing may be technically true (it can be defeated by certain types of mines, etc.), I wonder to what extent they may also not be terribly valid.

What I mean is, I would guess that armies that use high-quality, well-sealed mines, probably also aren’t the same ones that indiscriminately mine farmland in order to inflict misery on future generations of peasant farmers. Those ones would likely be using the cheapest mines available to them at the time.

And by the time a nasty civil war has calmed down enough that you’re contemplating how to demine vast tracts of farmland, it’s probably been years since the mines were laid, and you may know pretty well what sort of mines you’re dealing with and what techniques apply to them.

I dunno though – my closest contact with demining is that a friend of mine once did some, I think it was in Bosnia. I know nothing about it myself…

Dragonfrog puts it pretty well. This is for cleaning up the messes of the past, not for soldiers to navigate enemy mine fields.

http://en.wikipedia.org/wiki/IGEM

My colleagues who work on this have previously won a medal at iGEM for an arsenic sensor.

(Followed, of course, by squealing from the “frankenfood” nitwits.)

You know you’re all thinking it.

We are shown no pictures, no field tests, not even a hint that it has successfully worked anywhere.

If this does use GFP, then it would be very hard to make it work. GFP doesn’t “glow”, it fluoresces – it emits green light when exposed to blue light. And it emits roughly one photon for every hundred thousand you throw at it. So you would flood the area with bright blue light and look for weak green light (so it must be done in the dark). It could be done with the right filters, I guess. But instead of “glowing bacteria that finds landmines” this is “bacteria that express GFP upon binding molecules X, Y, and Z” – the post is misleading, although this is still an impressive feat of synthetic biology.

What strikes me as more likely is that the product of a transgenic protein turns green, giving the bacteria their distinctive color. That would likely be something similar to luciferase, but using another product instead of luciferin.

An interesting unintended consequence of this development, however, is that you actually may get more than just glowing biobricked bacteria is that they’re using plasmids as a transmission vector when growing bacteria. That means in theory they could pass on their plasmids to other bacteria that are similar enough to allow crossing over. Because biobricking uses common enough restriction enzymes, I could see other bacteria adopting this fluorescing reaction for other substrates as well.

Edinburgh University said the microbes could be dropped by air onto danger areas.

As positive as this development is, that quote is a bit sinister.

old landmines kill and maim lots of people.

The ones previously mentioned did not work with RDX landmines, only TNT.

http://herorat.org

Also, doesn’t seem anymore sinister than what can be accomplished with a giant payload of anthrax.

That’s not exactly the case. Some mines like the PFM-1 “butterfly” landmine are tightly sealed, and contain relatively small quantities of explosive. They can be buried in leaf detritus, so although they can be dispensed by air and are not strictly “buried” like we think of landmines, they present a considerable threat, particularly if they don’t go off. From this web page:

http://www.fourmilab.ch/minerats/figures/mines.html

note that there’s the PFM-1S, which has a 24-hour autodestruct mechanism; of course, these are never 100%, so unexploded ordnance (UXO) may remain.

Hands-and-knees probing is the only solution approaching 100%. Trusting lives to some wacky spray or a plant that supposedly changes color will never replace manual detection and removing. For those of us that don’t live in countries with a landmine threat, it’s easy to forget this. In areas where buried mines may be heavily overgrown by trees (Bosnia and Croatia, for example), these suckers can be really hard to detect.

And BioBricking: not as fancy as it sounds. To my understanding, it’s just a pre-built, user-friendly method of cloning. It’s like… just running a Linux GUI but not fiddling with the kernel yourself. You still use the command line, which seems scary to people who love Windows Vista, but someone else has vastly simplified the system for you. (I’m sure there’s a better analogy, but I’m on the bio side, so I don’t know enough about OSes. ^_^)