I have been thinking through the issue of debating this further and have decided to continue despite my earlier reservations (based on time commitment).
There are many issues at play here, so I will try to dissect them as best I can.
First, as to the issue of drug resistance in HIV and TB, there is a wealth of information --both from the past and current triple therapy regimens. A recent
news report covering a two year old study that examined compliance and the development of drug resistance indicates that compliance is the key issue in developing drug resistance and that drug resistance occurs in patients who are not properly compliant:
"Those who took 80 percent of their medication were likely to develop resistance most quickly. This probably means that they are taking enough medication to create a selective pressure," Dr. Harrigan said. "Which means that resistant virus mutations still have an opportunity to replicate instead of reducing the viral load to levels so low it can’t replicate at all."
The two key issues cited are those that I have harped on for the past two weeks – compliance issues that result in variable selection pressures (as opposed to constant pressure) and the strength of those pressures. This also speaks to the silly twisting of my arguments that Kleinman persists in perpetrating. Of course, if selection pressures are placed and permanently removed, information is lost. It is the irregular selection condition of poor compliance that provides a pressure but also allows sufficient reproduction that is important to the evolutionary process, especially as it relates to drug resistance.
When selection pressures are strong enough to decrease viral replication to the point where variability shrinks nearly to zero, then productive changes that will allow the virus to escape the selection pressures shrink nearly to zero. We all know that potent selection pressures slow the evolutionary process for the organism selected. That, in fact, is the definition of potency in a selection pressure (and why I have been harping on potency). The more potent a pressure is, by definition, the fewer offspring will emerge in the next generation. Fewer offspring means fewer experiments in variability, which means a slower rate of change. Triple therapy provides a profoundly (potently) selective pressure on HIV, as witnessed by significant reductions in viral load (I trust that I needn’t reference this phenomenon). But current triple therapy is effective only if compliance is nearly perfect (95+%) – if the selection pressures are kept in place and viral reproduction is kept to a bare minimum. When compliance is reduced to 80%, then enough viral replication occurs that new strains emerge and resistance develops. This level of compliance provides the same three selection pressures but still allows enough viral reproduction so that natural variability can produce resistant strains. So, by definition of selective pressures, 80% compliance with a triple therapy protocol constitutes three selection pressures that are less potent than triple therapy with 95+% compliance. The important factor is not the number of selection pressures but the potency of the selection pressures. The same phenomenon occurs with less potent HIV triple therapy regimens in which compliance was maintained at 95+% as I argued in a previous post and provided documentation for the emergence of resistant strains. Those particular triple therapy regimens are no longer used because the three selection pressures they employed are not potent enough to prevent the development of resistance. In fact, 95+% compliance on those regimens was the best predictor of resistance development. Once again, it is not the precise number of selection pressures that slows the process. It is the ability of those pressures to decrease reproduction/replication to a level where variability in the next generation is reduced.
Kleinman has repeatedly argued that three pressures act synergistically to reduce replication, and therefore variability, in HIV treatment. But this is not the case for all triple therapy regimens. As shown above, it is not the number of selection pressures that is critical. Three pressures will always slow the process more than one or two pressures – at the very least pressures must be additive, if not synergistic. But, three pressures will not necessarily stop the process or slow it to a point where “evolution becomes impossible”. The data from early trials with less potent triple therapies and relatively poor compliance with current triple therapy protocols demonstrate this fact unequivocally. It’s simply a numbers game. If enough progeny emerge in subsequent generations, then variability will, most likely, produce an organism that escapes the selection pressure(s), regardless of the number of those pressures. If the pressures are overwhelming, the species becomes extinct. If the pressures are so potent that variability is reduced to a minimum, then change will be slow or non-existent. If the number of progeny and variability are high enough, then new organisms evolve to escape the pressures.
With HIV triple therapy, if resistance to one drug emerges, the others can typically keep the viral load so low that variability cannot overcome the effects of the other drugs. It is not the case that three changes at once demolish the function of the target protein(s) – they proposed synergistic model. It is not possible that this is the explanation or multi-drug resistance in HIV would be impossible through whatever mechanism. But we do see multi-drug resistance in HIV, cited as evidence earlier (on more than one occasion and by more than one poster). Rather, no direct advantage necessarily accrues to the newly emerged strain resistant to one drug – it is still hit by the other two and viral loads remain low. If viral loads remain low enough, the process will be excruciatingly slow. If, however, viral loads increase due to poor compliance, then resistance to all three drugs becomes an option – and this is precisely what we see.
How does this relate to the ev model? I’m not sure that we need really ask the question simply because ev was never designed to model the entire evolutionary landscape. From my limited knowledge of it, ev was designed specifically to answer one question – can random mutation and natural selection increase information? Ev demonstrates that this simple process can do so without invoking the more profound process of gene duplication with subsequent modification. That argument is, therefore, laid to rest.
Kleinman insists that ev is an accurate portrait of the entire evolutionary landscape, but I don’t see how it is from the discussion I have seen. The population sizes are far too small to represent nature. This isn’t the fault of ev, though, because it wasn’t designed to answer the questions currently placed upon it. Recall that the single most important factor in the development of resistant strains of HIV is compliance. Poor compliance means more viral replication. The population size is absolutely critical to discussions of this sort. Reduce the population size enough and change slows dramatically. Adding three selection pressures necessarily reduces population size more than a single pressure of equal potency. So, necessarily, three selection pressures, holding potency of the selection pressures constant, will impact the speed of evolutionary change more than a single pressure. If those three pressures are modeled in a small enough population it will appear to stop the process cold in its tracks. There is likely some population threshold over which variability will almost always allow escape from multiple selection pressures, but I cannot pretend to know what such a threshold is – it would vary depending on circumstances and the organisms involved anyway.
Take an example run, in which the “genome” size is relatively large and three selection pressures are applied to an initial population size of 1000. If those criteria produce a “perfect creature” only after millions of generations, then why not the converse – millions of initial “creatures” may produce a “perfect creature” in approximately 1000 generations? How do we know this isn’t possible? Increasing initial population sizes to 10,000 or so will not help to elucidate the phenomenon because the variability is probably still too low. We simply cannot propose a linear relationship between increasing population size and numbers of generations to “perfect creature” over such a small range of population sizes. We would need to run simulations on these huge population sizes to tell, or do the commutative math and figure that a large population size is likely to produce an escape creature in short order. That is certainly what we see in nature and definitely what we see with HIV and TB. Maintain a low population size and triple therapy significantly slows the development of resistance. Allow population size to increase and resistance develops, even in the presence of the same three selection pressures (80% compliance condition). The above discussion, of course, neglects the significant problems with a “perfect creature” in this process, but I hope highlights the key issues.
I don't know if any of this discussion helps or is merely rehash of earlier debates in this thread (I suspect the latter), but here it is anyway. Since this is already too long, I will leave the rest until later.
