Reading the original paper, they compare Botswana and S Africa and find a correlation between seroprevalence and viral load. There is not good evidence to suggest this is causative however--this may be, for example, that as you get to higher frequencies of cases differently aged people are more likely to be infected and show different viral loads. A direct causal link between these two factors doesn't make a huge amount of sense, because (AFAIK) HIV is thought to sample all viral sequences within each host (so population-level selection should represent individual-level variation).
Sometimes if the parasite moves from host to host as part of it's life cycle, and this can include aiding the death of one host. An example of this is a parasite that causes frogs to grow extra legs so that they are easier to catch for birds. The parasite then transfers to the bird, where it can lay it's eggs.
Parasites can have multiple host species, some in which they are benign and some in which they are deadly. Malaria has been killing humans for thousands of years. Also for example, HIV.
"In theory, if we were to let HIV run its course then we would see a human population emerge that was more resistant to the virus than we collectively are today - HIV infection would eventually become almost harmless."
Because millions (or billions) of people would die, leaving only those with resistance.
> "Twenty years ago the time to Aids was 10 years, but in the last 10 years in Botswana that might have increased to 12.5 years, a sort of incremental change, but in the big picture that is a rapid change."
20 years ago, the ARV's and mixes weren't what they were 10 years ago and awareness + access to therapy also changed. Therapy initiation guidelines changed - which would directly impact time to AIDS.
Throw that statement away in relation to the meaning of the study results - except if this statement is there, it raises a concern about bias or mistakes in the study due to an insufficiently rigorous handling of the data and contexts/meaning.
No, not make them better, but there is an interesting story here which I heard when I attended a lecture at my job researching HIV.
HIV-1 is the "HIV" we all know and fear. While HIV-1 was derived/evolved from SIV (simian immunodeficiency virus), it's actually not as similar to SIV as HIV-2, another strain.
HIV-2 is far less dangerous and transmissible in comparison to HIV-1, and doesn't exist in large quantities outside of sub-Saharan Africa. Critically, there is some new research which shows that infection with HIV-2 first provides a substantial amount of protection against later infection by HIV-1. People who were infected with HIV-2 before being infected with HIV-1 had lower viral loads and slowed disease progression in comparison to people who were infected with HIV-1 alone.
This lead the researcher giving the talk to state the extremely stunning (yet perhaps true) fact: "HIV-2 is the best vaccine against HIV-1 that we have."
Sure it can - think of smallpox: the first smallpox "vaccine" was actually inoculation with cowpox - a similar, but less virulent virus. The provoked immune response rendered a smallpox infection less lethal, and provided a measure of protection.
You are not the first to wonder that! Back in the early 90s some scientists actually tried it on people with severe AIDS http://www.rci.rutgers.edu/~piecze/Lancet.PDF - some seemingly improved, but the trial was too small and preliminary to be conclusive. When they published those results, there was a huge backlash about ethical concerns and the research got shut down.
I took a class taught by Dr. Pieczenik about 5 years ago and he still believed this idea would work if it was allowed to proceed.
Surely this process means that HIV is just as capable of evolving into something even more lethal over time, given the right environment? Given the relatively closed environment of a country like North Korea (for example) where diversity is restricted, its evolutionary path may take a different route than that observed by the University of Oxford research team?
Can random mutations make HIV more lethal? Sure. The argument here is that lethality makes HIV "less fit" from an evolutionary standpoint, which makes sense, as, given the means of transmission, killing the host isn't a particularly useful thing for the virus to do.
> Surely this process means that HIV is just as capable of evolving into something even more lethal over time, given the right environment?
It's unlikely. HIV doesn't "want" to kill the host because that also kills the HIV, so it's the opposite of what it wants.
You would need an environment that favored killing the host over keeping it alive, and considering how HIV is transmitted that is unlikely. (Unlike Ebola where it can happen because Ebola is best transmitted from a corpse.)
It's like the foxes that over-populate, eat all the rabbits and then all die. The foxes would be better not to reproduce past the limits of their environment, but each individual fox passes on more of it's genes if it doesn't restrain it's own breeding. And so those genes propagate to the detriment of the community.
You need very strong group selection to restrain breeding, and this becomes less and less likely as the population size increases (so each individual has less and less affect.) See The Tradgedy of Group Selectionism: http://lesswrong.com/lw/kw/the_tragedy_of_group_selectionism...
Even when organisms are selectively breed by scientists with group selection to limit their population size, they often just become cannibalistic rather than restrain their own reproduction. Killing someone else's kids is better than killing your own.
Essentially the same thing happens with viruses, where over reproduction means killing all the hosts cells.
90-95% of adults had it at some point. Most people develop immunity to it and so it doesn't survive in them.
I'd say the most successful virus in humans would be the common cold, given how frequently we seem to get it, although it's more of a massive family of virii.
jbogp, I'd be interested to chat with you about your experience interviewing with DeepMind (found that searching HN for DeepMind info). My contact info is on my user profile page. Sorry about the thread-jack.
"Unintended" side-effect yes. But not really avoidable. The very nature of viruses means destruction, competition for resources with other physiological processes in the host, and they only have a limited capacity to tune their reproduction for optimal transmission. Too little reproduction (and in the wrong places) means too little transmission, too much reproduction means too much damage and thus also less transmission.
It's quite avoidable. The viral load in various ecosystems can be tremendous, and many are generally benign. Even the non-benign generally don't kill their hosts.
An interesting tidbit from xkcd's "What If" series: individual cell mortality from viruses in the oceans is about 20% per day.
Jared Diamond has offered related observations on pathogens and parasites in tropical vs. temperate climates. It strikes me that an ecosystem in which there are periodic resets (e.g., killing winter frosts, or long summer droughts) offer a more amenable set of living conditions to those species not affected by those resets, whether by adaptation or technology.
So mostly people are talking about virus which can be actually identified and studied. Up until very recently this was pretty much restricted to those who reproduce massively and leave at least some clues. We know next to nothing about benign virus in eucaryote hosts (or at least I don't).
Also, I don't think "benign" virus compete with pathogenic virus for anything. This may seem strange, but unless the organism is completely shutting down, and probably not even then, only an extremely small fraction of cells are infected by any one virus. That's why you still have a nose after surviving the flu. Any competitive mechanisms would have to be more complicated.
This negotiation is rather weak since it doesn't recognise the passage of time. Once the viruses have reproduced to their satisfaction, they can renege on the deal when humans have a weaker negotiating position.
That is to say, "Don't negotiate with terrorists."
Fortunately the human/mammalian immune system does not typically permit this kind of deal. With a few exceptions: including HIV.
If killing you means that the pathogen can spread more rapidly, it will evolve to do so. I don't see the HIV virus gaining anything by killing the host as it depends mostly on sexual intercourse to spread. Correct me if I'm wrong, but, Ebola would have been a better example in this case, I suppose.
> If killing you means that the pathogen can spread more rapidly, it will evolve to do so.
Sure, but killing the host is not generally the goal of a pathogen. Killing the host typically means the pathogen won't spread any further, and that direction leads to the pathogen's extinction. Or, the death of this family line, if you like.
But, before worrying about that, if a virus does not address the immune system, it will not be able to spread at all - there will be no family line. The initial goal of a new virus is to evade the (highly advanced) immune system.
Ebola apparently has a very naive way of bypassing the human immune system. That is, overwhelm the system, so that the host dies, but you have a small chance of spreading. This works, sometimes, for a little while, but not too much, maybe just enough to gain a few generations of evolution to try and improve your strategy. If you are already a latent virus in a related species, you might get a few attempts at this, as Ebola has.
HIV does this in a frighteningly more effective way - a way which is more evolved - hide from the immune system while maintaining a route to spread to other hosts. This has to be a very subtle balance - hiding from the immune system is no mean feat. This is why I used HIV as an example - HIV somehow negotiates its own and the host's survival with the immune system (where Ebola has not been able to).
The dance between immune system and virus ultimately leads to AIDS. Death of the host is not ideal for the local strain of HIV, but it had a good chance to reproduce in other hosts, so it's a good result for HIV as a whole.
To become an effective and latent human virus, you need to evolve some sympathetic way of defeating the immune system that doesn't have a significantly negative result for the host. It's likely that HIV is still on that path, and Ebola might be on the start of that path.
This is why we observe viruses to become less deadly and more easily spread over time.
It's not necessarily true that this is the only evolutionary path, but the immune system presents such a huge barrier that you have to overcome this before meeting any secondary goals.
You are correct, Ebola's high mortality rate is a by-product of how it spreads.
Ebola virus is successful at spreading due to the fact that it's in various fluids secreted by endothelial cells (which are everywhere in the body). The virus has mutated to let those tissues start to "leak", which means a person infected with Ebola is more likely to leak fluids (which is why it sometimes causes bleeding out of mucous membranes). Since there is an increase in the endothelial fluids released, there is more exposure to those trying to treat a patient with Ebola.[1]
The way the virus spreads (increased discharge of endothelial fluids) is what causes it to be so fatal, which is a by-product.
>If killing you means that the pathogen can spread more rapidly, it will evolve to do so.
The plague/black death/Y pestis is another is a great example of this. It keeps its rat host just alive enough for just long enough for maximum use of the rats body. The rat then dies, which sends sends it plague-ridden fleas off to infect other rats.
While this is a communication pattern, it is still governed by evolution. Therefore, as a host, you get one vote.
Either you live at least long enough to spread the virus a little bit, and you vote that what the virus is doing is good, or you die before spreading, and you vote what the virus is doing is bad.
No, that's just evolution - plain, simple and extremely dumb. Viruses don't really "want" anything, they are just biomolecular machines that run around replicating themselves and screwing things up; some manage to copy themselves, some don't - and this way, generation after generation, you're left with those viruses that are good at copying themselves in the environment they're in. Evolution is a statistical phenomenon.
In other settings, a simple approximation of what Evolution does will balance your portfolio and find Nash Equilibriums for zero sum games much more quickly than linear programming (query: multiplicative update evolution).
Having to wait for things to die can be restrictive (but genomes are declarative, what does that buy?). Roughly, with what amounts to the assumption that Lamarckian evolution is more or less not true, you can characterize evolution's learning power as almost as good as PAC (query: statistical query learning evolution).
Except that, depending on your definition of "intelligence", evolution even as a statistical process fits most to all of the criteria necessary to be intelligent. Basically any definition of intelligence that fits the stuff you see in the movies will also tend to fit evolution. "Has goals, biases set of possible futures toward completing those goals, demonstrates creativity", and so on. It's just that evolution, as an intelligent entity, is utterly, unbelievably alien - psychologically, physically, and 'physiologically'. Which I guess makes sense, right? Aliens aren't going to have rubber foreheads and five fingers; they're going to be utterly unbelievably alien.
1. You're right, our intelligence is the same thing. Intellect is a statistical creation that makes us good at surviving.
2. But usually, when people talk about our intelligence and our having goals, they aren't talking about the fact that we statistically exhibit tendencies towards self-preservation (which is what we're talking about viruses doing). It's that we literally have conceptual goals, which can be written down ahead of time and talked about.
aka you won't find a virus with a to-do list. Conflating that intelligence with viral 'intelligence' is kinda disingeuous.
--
That said, I do agree that we might/will find alien life that looks nothing like what we expect and has 'intelligence', for some reasonable (non-virus) definition.
I would not be inordinately surprised if we managed to find this in the motion of waves in the ocean, or in the swirling gases of the sun, or the gravitational interactions of the asteroid belt, or in tremors in the earth, or in hive-minds of billions of ants.
And I definitely reject dualism and fully expect that a human brain simulated in a computer would be as intelligent as we are.
I'm not sure what the criteria for an 'intelligence medium' is, but it seems to require sufficiently high entropy to allow for complex patterns to form, but not too high - there has to be enough order and especially a not-too-high temperature such that self-contained systems can form and replicate without being blurred back into the whole system.
True, but it's a kind of a "null garbage collector" case. It's the dumbest type of intelligence, a random one. But it did also created intelligences less dumb than itself, if you view the evolutionary process as separate from the workings of human mind.
Obviously viruses are not intelligent and they don't "want" to reproduce.
However, they do reproduce, just like pretty much every other form of life. The question of why life does that is closely related to the question of what life is and how it came to exist. Those are deep questions and we don't really have a clue as to the answer right now. There's certainly nothing in the fundamental physical laws of the universe, as we understand them today, that seems to indicate or require the existence of life.
Closer than you might think. Howard Odum in 1971 suggests that humans are something of a decomposer, working on a backlog of accumulated energy and mineral deposits:
Sometimes in half-seriousness we say that man may have been evolved by the system as a mechanism to get the fossil fuels and other minerals back into circulation. We hope he is pre-adapted for other roles after that.
From Environment, Power, and Society.
An analog I use is to opportunistic benthic whale fall communities. The emerge and develop rapidly. They do not sustain.
Is life an accident, or a predictable outcome of physical laws? That's a pretty fundamental question, and one we should be able to answer scientifically. But we can't right now.
Fundamentally life is just patterns of information and their associated support structures capable of reproducing and adapting to their environment. There really is no answer to the question of "why" life reproduces because life that does not reproduce ceases to be life. The reproduction component is inherent in its definition.
As to why our universe is one which supports life, nobody really knows. Some invoke the anthropic principle, but that is more of a philosophical answer than a scientific one.
> As to why our universe is one which supports life, nobody really knows.
There's an implicit distinction in this statement, between the universe and life. Really, life is part of the universe too.
I'm familiar with the anthropic principle, but it doesn't tell us whether life is a necessary outcome of the physical laws of our universe, or whether it's just one of many accidental outcomes that are not forbidden by the laws of our universe. Both of those situations satisfy the anthropic principle, but are very different.
For example, we could speculate about a universe in which the nuclear strong force is stronger than it is here. Based on our theories of quantum mechanics and gravity, we can predict how that would affect, for example, star formation. Isaac Asimov explores this idea in "The Gods Themselves."
But Asimov also predicts what alien life would be like in that universe, and there he's on much shakier ground, because nothing in our theories of quantum mechanics or gravity predicts life. We don't know how chemical systems become alive...there's even not solid agreement on the definitions.
I thought the question of why life reproduces is easy to answer: things that reproduce become greater in number, and things that reproduce better become even greater in number than things that reproduce worse. Thus, the world is filled with things that reproduce really well.
