1. Patients launch $1.27 million crowdfunding campaign for ME/CFS gut microbiome study.
    Check out the website, Facebook and Twitter. Join in donate and spread the word!

Broader Perspective

Blog entry posted by anciendaze, Jun 28, 2011.

This post is different from previous ones I've made. I'm going to try to put current problems and controversies into a broader biological perspective. This will not appeal to everyone. The problem I'm grappling with is a fundamental one with implications far beyond the moment. Abstractions mentioned are likely to be unfamiliar. This does not mean they are vague, sloppy or inapplicable, only different.

Darwin struggled with a central problem of biology which he rightly used in the title of his famous work 'the origin of the species'. Considering the state of the art at the time, he did a remarkable job of demonstrating that random variation and natural selection were observed facts. In the end, however, he had trouble with the apparent existence of fixed species on many time scales. This doesn't say he was wrong, only that theoretical biology was incomplete.

In the 1970s, a radical thesis called 'punctuated equilibrium' came into vogue. By that time there was sufficient fossil evidence to show that many species undergo remarkably little change for long periods of time, until they disappear. This was, in a sense, a general observation rather than a theory of how this came to be. If there was a comprehensive underlying framework, it passed most biologists by. You can get a remarkable range of explanations from various biologists today.

The revolution in molecular biology was just getting underway at that time. If anyone proposed a human genome project back then, it could be demonstrated to be absurd. Since that time there has been a great deal of work, but if there has been an 'origin of the quasispecies' it has escaped me. This is what would be needed to explain the evolution of retroviral sequences.

We see a situation in the discussion of viral species which can easily explain diffusion away from an ancestral species, but has to rely on near miracles or vast times for the origin of those species. This is particularly apparent in the debate over the origin of XMRV. The problem is much more general.

When researchers chase contamination, their focus is very narrow. They aren't thinking about general biological principles. Most of the time they aren't able to trace the exact path by which contamination occurred. Even in successful cases, the story is unlikely to get written up and published. It becomes part of the folklore of the field. Assumptions are unlikely to be exposed to critical reappraisal.

One principle of large scale evolution is that convergent evolution never produces identical species. You may find a marsupial small-animal predator that resembles a fox, but detailed examination will always reveal a huge catalog of differences. It might have a jaw that works like a fox's jaw, but the bones forming it will have a different embryological origin. You can't simply move the jaw from one species to an unrelated species.

At the level of molecular evolution of quasispecies, things are less clear. Recombination is not limited to the same species or genus. Convergence to a particular sequence can take place in a specific environment.

We can talk about the 'fitness landscape' in a particular environment as if it were a kind of topographic map. The different coordinates would correspond to different sequences, while the altitude would give you the likelihood that sequence will replicate well in that environment. Small-scale evolution can then lead to hill climbing, which ends at a peak. Many similar sequences may converge on the same peak.

In the language of dynamical systems the altitude is reversed. Now you replace hill climbing with a process like water running downhill. The region leading to a particular common end point is called a basin of attraction, and the end point is called an attractor.

Mathematically, this can be modeled as an iterated map which converges to a fixed point. The property of having a fixed point is very general, but it is typically true of non-linear maps. Most of the convenient mathematics taught in schools is linear.

A fixed point is not the only possible outcome. Another possibility is a limit cycle, where the same sequence of states repeats endlessly. There is even a stranger possibility appropriately called a 'strange attractor'. (Poincar type-three motion for high-brows.)

At this time, I only need to consider a fixed point. In a very simple environment, like a carefully-tended cell line, the fitness landscape can be simple, dominated by a single peak/attractor with a large basin of attraction. In this case random changes will destroy many sequences, but a sequence with a rapid rate of replication will likely come to dominate. It doesn't need to defend against predation, because there is none in this environment. It doesn't have to grub up nutrients, humans will supply them. All it need do is replicate as fast as it can.

What I'm suggesting here is that a sequence like VP62 is a fixed point in the environment where culturing takes place. If any one of a range of sequences are present in the original sample, over time these will converge on VP62. Change the environment, and you may get VP35, or the virus contaminating 22Rv1.

This is only possible if there is a sequence in the basin of attraction leading to those fixed points to begin with. It will also require some number of cycles of replication to perform the small-scale evolution required to converge. A highly-specific test for a particular sequence will initially fail. After some number of cycles of replication, it may succeed. This does not require a miracle. It should be expected to happen.

This fits a number of observations in the present controversy.
anciendaze

About the Author

As the name suggests, I am old and dazed. The avatar illustrates my rule of thumb: "Hang on! This ride isn't over."
  1. Marco
    anciendaze wrote :

    At the level of molecular evolution of quasispecies, things are less clear. Recombination is not limited to the same species or genus. Convergence to a particular sequence can take place in a specific environment.

    We can talk about the 'fitness landscape' in a particular environment as if it were a kind of topographic map. The different coordinates would correspond to different sequences, while the altitude would give you the likelihood that sequence will replicate well in that environment. Small-scale evolution can then lead to hill climbing, which ends at a peak. Many similar sequences may converge on the same peak.

    Thinking about the 'fitness landscape' favouring convergence to a particular sequence, its is hard to conceive that the in vitro environment of cell lines would have much in common with the environment of pathogen/host interaction in 'the wild'.

    It just occurred to however that I was comparing and contrasting cell line development with the wrong other environment. What we need to compare are the environments in which XMRV has been detected and sequenced in cell lines and from patients. Rather than very different environments there is a great deal in common between the xenograft/repeated passages through mice cell line development and the process of patient samples being cultured and subjected to iterative rounds of cloning/PCR. This amplification process may actively favour convergence to a particular sequence 'reference XMRV' in a similar way to cell line culturing; all but eradicating variation found 'in the wild' and hence feeding the contamination meme.

    Which of course would then present a considerable problem in detecting and sequencing variants present as a natural quasispecies in infected patients if the very process needed to amplify and sequence them eradicates variation.

    Its fun to speculate of course but not without some basis in observed facts; e.g. :

    Identical retrovirus sequences appearing in distinct murine cell lines :

    Viral RNAs from three cell lines, SL3, 4, and 7, were indistinguishable from one another.

    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC256722/


    convergent evolution of host restriction factors :

    Third, there is a striking example of convergent evolution at the TRIM5 locus where a new hybrid TRIM5/cyclophilin gene has independently arisen via retrotransposition and has been selected in two distinct primate species [22][27].


    The possibility that the culturing/PCR process itself provides a similar environment to cell lines :

    The libraries of recombinant retroviruses contained many hundreds of members that were subjected to the functional analyses outlined below. However, sequence analyses of these libraries revealed many discrepancies between the amplified sequences and sequences present in the genome databases. In principle, these discrepancies could arise as a result of (i) natural variation in endogenous CA sequences, (ii) sequencing errors in the genome databases, (iii) recombination between closely related amplicons during PCR/cloning and (iv) Taq polymerase errors. Thus, the ensuing analyses exclude CA-NTD sequences that did not perfectly match a variant that was present in the sequence databases.


    http://www.plospathogens.org/article/info%3Adoi%2F10.1371%2Fjournal.ppat.1000181
  2. alex3619
    We are likely looking at a viral swarm which only reveals itself when one member reaches such an optimum. It is premature to say how many other similar sequences are present.

    Hi anciendaze, these sentences sum up my point of view currently. The current research is looking at specific instances rather then complex system dynamics. Almost everyone seems to be doing that, its how the disciplines have developed.

    If a researcher is looking at the swarm in one environment, they might find dominance of a specific viral strain. They then name it or find its existing name. Another researcher looks at another environment, finds a similar swarm, isolates the dominant virus, sequences it, then finds or gives it a name. The two can then get together and debate on how similar or different the two strains are. The swarm is ignored.

    We currently still don't have the technology to really study viral swarms. We are forced to rely on points of data for a complex multidimensional situation. What we can look at is the totality of all the data points and ask if there are alternative explanations.

    Bye
    Alex
  3. anciendaze
    ...I wonder are there instances of 'spontaneous' occurances of retroviruses or host responses in cell lines and/or xenografts in the literature? These will likely be reported as contamination. The tacit assumption in the literature of natural diffusion away from ancestral species, with no consideration of convergence to a fixed point in a given environment, almost dictates this.

    It is a standard observation in many fields that a large change in scale is tantamount to a qualitative change. By normal standards, elephants and fleas live on very different scales. Their evolutionary scales are very similar, the same principles apply.

    Quasispecies of sequences are on a dramatically different scale. There is the additional complication that recombination can 'mate' things which are remarkably different. This kind of change demands reexamination of assumptions.
  4. Marco
    I need to reflect on (as in take some time to get my head around) your last post.

    Iin the meantime, odd that from what I have read of virology, they seem to be concerned only with random events followed by entropy while apparently ignoring the inevitable and necessary impulse to increased order that characterises any complex system.

    I wonder are there instances of 'spontaneous' occurances of retroviruses or host responses in cell lines and/or xenografts in the literature?
  5. anciendaze
    Marco, I am not positing any strange attractors at this time, only the existence of fixed points in particular fitness landscapes.

    You also need to realize the assumed metric for the space where a normal distribution forms is likely to be equivalent to a Euclidean metric. Legendre and Gauss were looking for measures which would allow them to use least-squares methods when they began reasoning about such distributions. This naturally leads to linearity assumptions, and linear mathematics has a much easier time explaining why things fall apart than explaining any form of convergence.

    (You can check on models of wave packets in many fields. In electrodynamics, you run into the same problem with time symmetry. The mathematics includes an 'advanced potential' which must be removed on the grounds that it violates causality. The same tendency to fall apart shows up in the time domain when electrical engineers design ideal filters. They have to be careful to use only 'causal' filters in designs of physical devices. Acausal filters could respond before a signal reaches them. The model is not "das Ding an sich".)

    Just to clarify one aspect I did not dwell on, while there need only be one sequence which mutates into the fixed point form to dominate a culture or cell line, this does not have to be the same one in different contexts. We are likely looking at a viral swarm which only reveals itself when one member reaches such an optimum. It is premature to say how many other similar sequences are present.
  6. Marco
    Another very interesting post anciendaze.

    I'd been thinking about the possiblity that, in the context of a quasispecies infection (or the presence of a variety of murine/human sequences in vitro) that XMRV might represent 'the most likely replication competent retrovirus sequence' to emerge in vivo or in vitro and that this is a continuous rather than unique occurance.

    I hadn't considered the process in terms of a basin or basins of attraction but more simply as a normal distribution of replication competetent varieties produced through recombination, with XMRV representing the most frequent (and prototypical) sequence with more distantly related sequences occuring much less frequently at the tails.

    Prototypical or 'reference' XMRV would be much more likely to be detected in patients and in cell lines and easily dismissed as contamination whereas it may in fact represent a marker of a quasispecies infection.

    Positing a strange attactor complicates interpretation enormously. If recombination is a continuous process then you have the potential for sequences to be undergoing point mutations and diverging from the 'ancestral' form as per the standard model of diversity while simultaneously XMRV like recombinations are constantly being created anew with the tendency over time to more and more like the protoypical XMRV sequence dictated by the basin of attraction. Phylogentic trees become meaningless in this context.

    Questions remain whether in vitro and in vivo represent the same 'landscape' and if there are one or many possible 'solutions'.
  7. Enid
    Yes indeed I meant to add at the viral level things must move rapidly. And part of us so research in this area as we see (with ME) a very complicated field. Thank you for the information on pygmy marmosets - delightful.
  8. anciendaze
    Don't you recognize me?

    Those are called "finger-monkeys" in common parlance, though they are officially several different species of pigmy marmosets.
    http://www.buzzle.com/articles/finger-monkey.html
    http://scienceray.com/biology/the-ti...finger-monkey/
    http://www.google.com/search?q=pygmy+marmoset&hl=en

    The one in the picture illustrates my rule of thumb, "Hang On! This ride isn't over."

    The problem with studying the evolution of viruses is that they can evolve faster than researchers.
  9. Enid
    anciendaze not such a deep thinker as you. But much appreciate Darwin (visited his home in Kent and noted a loo by his desk - GI problems - and all the devestation for him of his discoveries being brought up "orthodox"). I also recall at the turn of this century noted biologists declaring "this is the century of the virus" - and so it seems now of which ours will hasten understanding. It's nature to evolve - bug or otherwise - so I think researchers must understand that. It is life and survival (best adapted).

    What is that delightful animal on your thumb. ?