It was once thought that certain diseases (such as syphilis) could be cured by passing the disease on to someone else. This is wrong, of course, but what if it were true?

How:

The disease would be a virus that operates much like a chain letter or pyramid scheme. The virus' DNA would contain a few "slots" where it would record the genetic identity of the last few hosts it had been through (by extracting enough "junk DNA" to identify an individual). When the virus encounters a new host it has never seen before, it adds it to the list. When the virus encounters a new host that it *has* seen before, it goes into a special "shutdown" mode where it signals any other local copies of the virus to cease replication (but probably remain latent).

So, in order to activate the "shutdown", you have to infect someone else, and then have them re-infect you to collect your "referral" bonus. (Or maybe you have to do this with 2 people, to ensure exponential dissemination.)

(Additional mechanism is needed to make sure that the new individual wasn't already infected, etc.)

Why:

From the disease's "point of view", evolutionarily speaking, the advantages are obvious; the disease is enlisting the full consciously applied resources of its host to spread itself! Imagine the public health issue this would raise as the infected are actively seeking to break quarantine to infect others. ([Monkfish], at the right, offers several reasons why such a virus might not have evolved naturally despite the advantages.)

From our point of view, of course, this would be a disaster. Engineering such a disease might be a good "evil freedom foundation" idea.

Might make a good science fiction story.
-- egnor, Jan 01 2003

There are some reasons apart from complexity why this would never have evolved on its own. The requirement that its hosts understand how it operates rules out animals (the source of many popular human viruses) and humans themselves down until very recently -- widespread knowledge and acceptance of the principle of infection itself is very recent, and this is more complicated yet.

The "shutdown" part may have some problems. Effectively, reinfection is with a new strain of the virus. The new strain might recognize the host and shut itself down, but it also needs to deactivate the original strain. There may be a mechanism that would make this possible, but it seems like a lot to ask of a virus.

A possibly more important problem is that evolution would favour strains of the virus that cheated by refusing to go down when when the host fulfilled its end of the bargain. I'm not certain of the logic here, but the scheme might quickly collapse, leaving you with a conventional virus.
-- Monkfish, Jan 01 2003


You make good points.

It's true that this presupposes a degree of intelligence on the part of the hosts, though I think even before the acceptance of the germ theory of disease it would have had some success.

It's also true that the "shutdown" requires a signalling process of some sort, which is not only complex but subject to spoofing; if I were a human researcher fighting this disease I'd focus there first to see if I couldn't fake the signal. I don't think the signal itself is impossibly difficult, however. A relatively straightforward signal would be to start injecting marker DNA into host cells that rather than producing new viruses (and killing the host) instead acts as a block against new viral infection. Eventually the targeted host cells will all be marked or dead, and the host itself will be cured (or dead, but that's the risk you take).

The "cheating" problem is perhaps more severe. If such a strain did arise, it would be in the "interests" of the original strain to distinguish itself somehow. It could even enroll medicine itself in this cause; if you can imagine a level of medical technology capable of identifying viruses but not of curing this particular one, then there might be tests that would tell you whether you have the version with the shutdown or not, and you could behave accordingly.

That's not so impossible to believe; after all, we have highly accurate tests for various strains of HIV but no cure. This *does* presuppose modern medicine, of course.
-- egnor, Jan 01 2003


I agree with Monkfish. The additional deactivation part of the virus will cost the virus energy during reproduction for little advantage, since most hosts will not be able to tell the difference between a close relative and the virus itself.

If you moved this the the field of computer viruses then mutations would not occur. There would be no close relatives (unless someone coded an imitator). The resulting human behaviour of those with infected PCs might be interesting to a game theorist.
-- st3f, Jan 03 2003



random, halfbakery