To Hugh and the peloton discussion group,
I did a little riding in spandex and have a small sense of the dynamics inside a peloton. I always thought it a marvelous experience while participating. The most striking oddity for me about the discussion is the focus on the group concept. Personally it appears an artifact of the human propensity to organize random events into a pattern that may or may not have any meaning, it sounds a lot like an argument about what the shape of a cloud appears to represent. Each individual argues tautologically to support their specific perspective. The way a peloton transforms over time exchanging members and breaking into subunits then reassembling seems to contradict any notion of group. A group definition dependent on time velocity or orientation is simply an instantaneous attribute and I think misleading. Hugh's model should display behaviors that surprise everyone simply because we are trapped by personal perspectives and expectations. A peloton is a merciless entity disposing of riders as callously as wind strips leaves from trees in autumn. Yet occasionally it appears merciful, a weak rider can be hidden in the shadow and carried along until it has healed, its only function to tell jokes. Sprinters are typically delivered to the attack position in such a manner. Often the sprinter in tow is not even related to the mules. The mules or pullers try every tactic possible to rid themselves of the unwanted hanger on. So affinity is not a parameter for apparent cooperaton. That kind of blows the "Genefur" concept out of the water. To the outside observer cooperation seems dominant, to the rider it can be either a pleasant exercise or one filled with dread depending on how close is the finish line. Some pullers will invite a good joker along for the ride knowing full well that he will be dumped when things get serious. I personally have dumped riders who did not do their laundry regularly enough! Yes I pulled ahead and hit his front wheel, Not Nice I admit but he reeked and I could get away with it or so I thought. I had to confess to Friam, how peculiar! The fixation with sorting based on speed, reserve energy etc. will lead to finding the best rationale to explain that idea. I just wish to insert a new dimension namely that the purpose of a peloton might be to deliver sprinters. Energy optimization is not the goal but simply a tool. The peloton is an artifact of a few cyclists with temporarily synchronous goals. The rewards in the peloton are human constructs, sprinters are evaluated by the trophy and pullers by delivering sprinters. The puller has only to select sprinters with the same logo as he wears, he forces the role of sprinter upon the one they identify as being the most explosive on that particular day. The pullers define the structure of the peloton and determine who becomes identified as a sprinter. The pullers are ruthless, and they rule the roost so to speak. Pullers can be absolute tyrants and discourage any individualism when they sense the target. They amuse themselves with little attacks at various stages to evaluate the strength of each little team. The sorting by velocity is a dangerously biased view of a peloton. It has some merit but underestimates or oversimplifies what might actually be happening. Let me put it in another context, a peloton is one thing to outside observers and another to the individual riders, it is neither and perhaps it is both. The model that Hugh builds will be watched closely as Hugh seems to be onto something very important. Let the model speak for itself and later we will all argue about the wonderful patterns that appear in the "clouds". Peloton strategy sessions following a race are intriguing since they typically have both a personal and collective viewpoint. The personal is always the must relevant. Good team leaders play the riders like chess pieces and have no room for sentiment. If Hugh has time to spare could you please let us lurkers know of some of your progress building the model. I am just getting started myself and am interested in "Group Dynamics" as artifacts of an external perspective. I hope Hugh has an easier time understanding the peromyscus peloton than was my experience with the local FOG ( Fast Old Guys) riders, I never made the grade!. My guess towards building a group model of sperm is to program the agents as Tyranical Pullers willing to destroy themselves and everyone around themselves for a programmed goal, the program would be moderated by proximity to the goal. Perhaps they start off simply looking for a shadow behind a stronger rider, when the shadow is inconsistent they switch to a stronger lead wheel, there is no group selection just individual selection for an easy ride. The individual monitors his own energy state and tries to keep expenses at a minimum. He does also recognize is own affinity group based on some markers and attempts to keep them close (just how close is Hugh's problem) In cycling it is possible to dispose of a follower by simply slowing your pace and letting your back wheel touch the following front wheel. Such dirty tricks are not uncommon. The follower almost always loses control. If not, he will return with a major attitude ! In this case the goal is not to win but to eliminate the opponents. One has to be careful when doing this or the falling rider might take out a group of your affinity clan as well! On the other hand this may be the function of some sperm to eliminate competition from behind. I will follow your progress Hugh, as you build your model , with much enthusiasm. I just pumped my skinny tires up and hope to do some lazy riding as spring arrives in Winnipeg. Dr.Vladimyr Ivan Burachynsky Ph.D.(Civil Eng.), M.Sc.(Mech.Eng.), M.Sc.(Biology) 120-1053 Beaverhill Blvd. Winnipeg, Manitoba CANADA R2J 3R2 (204) 2548321 Phone/Fax <mailto:vbur...@shaw.ca> vbur...@shaw.ca -----Original Message----- From: friam-boun...@redfish.com [mailto:friam-boun...@redfish.com] On Behalf Of Hugh Trenchard Sent: March 29, 2010 10:42 AM To: ERIC P. CHARLES; Nicholas Thompson Cc: friam@redfish.com Subject: Re: [FRIAM] Sperm pelotons; article in Nature Thanks Eric for taking the time to look through my post. For Nick's last post, I am not entirely sure what a "genefur" is, although it sounds like it is a reference to an inherent genetic trait, as you also discuss. Yes, I agree it will help my argument if I hone in more closely on what I mean by fitness, and I will add some description to clarify this. My useage relates to inherent physical fitness in terms of maximum power output capacity. That too needs fine-tuning because I refer to "maximum sustainable output", which is not the same as absolute maximum power output, and I would need to outline more carefully what this means. Regardless, I think there are ways of testing for the actual power-output capacities of individual sperm - I have seen references in the literature to testing procedures for this. Because I know very little about genetics, for my part I would be treading dangerously to begin describing the process in a gene-related sense (and I would not want to get into discussion about chromosomes), but to address the issue you raise (if I understand it correctly), it would be necessary to measure the power output of the sperm of individual male mice to determine the range of their output capacities and/or the sperms' average output. This is no doubt not easy, but I imagine there would be some sampling size that would provide an accurate indication of the overall output range. And certainly one would want clearly to correspond average sperm outputs and ranges with the genetic descriptions of the various mice tested, but this could be done according to a replication of the Fisher and Hoesktra procedures. It would also be necessary to determine percentages of energy savings that occur when sperm are coupled (if this does in fact occur). My model assumes that there is a difference in the average power output of individual males' sperm, whether related or unrelated or of the same species or not - a difference sufficiently significant to demonstrate that sorting occurs according to fitness (in the power-output sense) and not according to some mechanism for identifying the genetic relatedness of the sperm, as the authors of the Nature article appear to suggest. The fact that sperm aggregate indicates coupling and energy savings, which is why (in my view) the peloton model applies. In terms of chance, it seems to me Fisher and Hoekstra have taken a lot of care to establish that there is sorting beyond chance, but implicitly ascribe that sorting to some sensory/perceptual capacity of the sperm to identify related sperm. My model begins with their proven result that there is sorting beyond chance, and asks whether there is some sorting mechanism involved other than an unidentified mechanism to perceive the location of related sperm, which is intuitively problematic because (it seems) sperm do not have a sufficiently developed sensory system (i.e. eyes, ears, or other) to do this. My model provides a simpler explanation for the sorting process than the Hoekstra & Fisher explanation, because, in my model, sorting occurs according to self-organized energetic principles, and not according to a perceptual/sensory mechanism, as apparently implied by the authors. I can see how a basic computer simulation would be helpful as a starting point for making predictions according to my model, which I see is really my next step. Does anyone know how/where one could apply for some funding to resource such a simulation? I could develop it myself (and have developed at least one simulation, but it really needs to be worked through again), but it would happen a whole lot faster if I could engage someone more adept at computer modelling than me. ----- Original Message ----- From: ERIC <mailto:e...@psu.edu> P. CHARLES To: Nicholas <mailto:nickthomp...@earthlink.net> Thompson Cc: Hugh Trenchard <mailto:htrench...@shaw.ca> ; friam@redfish.com Sent: Saturday, March 27, 2010 2:54 PM Subject: Re: [FRIAM] Sperm pelotons; article in Nature Hugh, Very interesting model! One of my doctoral adviser's, Jeffrey Schank has demonstrated repeatedly that scientists are very bad at predicting what 'chance' looks like when trying to do experiments involving synchrony. This seems one of those situations, and the only way around it is modeling. Nick's sarcasm aside, he has a point, and it has to do with some of the flavor text surrounding your model (for geeks of the wrong variety to know what flavor text is, see: http://en.wikipedia.org/wiki/Flavor_text). If I can take a shot at identifying the problem: Rather than looking at 'fitness' as if it were a unified trait, you have created a model that needs some mutli-stage selection language (the better term escapes me at the moment). The reality is that what makes a 'fit' sperm is not necessarily what makes a 'fit' organism. To fix the flavor text of your model, you would need to explicitly recognize that (if the sperm sort, then) the sperm are going to sort based on a similarity in the genes that 'build' the sperm. Their sorting will be completely independent of all the other genes, or of any role that the sperm-building genes might later play as body-building genes. Ignoring chromosomal linkages (which you shouldn't), two sperm could be identical on all the genes important for building sperm, but completely different in terms of all other genes. Your model would thus al! low a much clearer test of the prediction that sperm identify each other in some way. It does so because it provides a vastly improved predicted relatedness due to chance. GIVEN: We would expect sperm to cluster along the race track based on the similarity of certain, specifiable genes. MODEL: If we know the genes important for building sperm, we can model the expected relatedness of sperms within a cluster. IF: Sperm are implementing some weird sort of kin selection mechanism - THEN: we would expect the relatedness to be significantly larger that what our model predicts. Any help? Eric On Sat, Mar 27, 2010 01:36 PM, "Nicholas Thompson" <nickthomp...@earthlink.net> wrote: Hugh, Even if it has nothing to do with sperm it is a nifty model. There is an idea lurking here that i dont know whether it plays a covert role in your thinking or not, but what about the fate of a "genefur" peletonizing. My email program is misbehaving and my computer is about to crash so I wont say more, now. Nick Nicholas S. Thompson Emeritus Professor of Psychology and Ethology, Clark University (nthomp...@clarku.edu) http://home.earthlink.net/~nickthompson/naturaldesigns/ http://www.cusf.org [City University of Santa Fe] > [Original Message] > From: Hugh Trenchard <htrench...@shaw.ca> > To: <nickthomp...@earthlink.net>; The Friday Morning Applied Complexity Coffee Group <friam@redfish.com> > Date: 3/27/2010 10:54:41 AM > Subject: Re: [FRIAM] Sperm pelotons; article in Nature > > Thanks for taking a peek at my post. Great que! stions, and they help me to > see how/where my descriptions can be clarified. > > On the paradox part - that is one of the really interesting features of a > peloton: the energy savings effect of drafting narrows the range of fitness > between the strongest and weakest riders. In contrast, think of a pack of > runners of varying fitness levels. There is negligible drafting effect - > there is some, esp if running into a headwind, but overall it's small enough > that it can be ignored for this illustration. Say there are 50 runners, all > separated incrementally by 1% difference in fitness; say they run a couple > of miles. If they all start off slowly at say the max speed of the slowest > runner, they can all run in a big group, separated only by enough distance > between them to keep them from kicking and elbowing each other. As they > pick up speed, the gr! oup thins into a line and are separated incrementally &! gt; by d istances that correspond to their differences in fitness. In the space > of two miles, they all finish individually in a single long line according > to their fitness, and it can be predicted accurately where runners will > finish if you know their starting levels of fitness. > > This is not the case with a peloton. For example at 25mph, riders can save > at least 25% by drafting (approx savings 1%/mph) - all the riders who are > within 25% fitness of the fastest rider can ride together even at the max > speed of the strongest rider. So their fitness levels are effectively > narrowed, and they can all finish together as a group (ie. globally coupled > by finishing within drafting range of each other), and so the paradox. Part > of the paradox is also that, while fitness levels are effectively narrowed > by drafting, it means, conversely, that a broader range of fitn! ess levels > can ride together in a group, which maybe isn't something that is clear from > my initial post (though it is certainly implied). Also, there are other > important things going on in a peloton which precede the sorting of riders > into groups, some of which I see I do need to clarify to make my model > clearer. > > Of these, particularly important are 1) the occurrence of peloton rotations, > and 2) points of instability when riders are forced into positions where > they do not have optimal drafting advantage. Below a certain output > threshold, when all drafting riders in a group are sufficiently below max > output, riders have sufficient energy to shift relative positions within the > peloton, and in this particular phase, a self-organized rotational pattern > forms whereby riders advance up the peripheries and riders are forced > backward down! the middle of the peloton. However, instabilities in pace > oc cur along the way, caused by such things as course obstacles, hills (when > lower speeds reduce drafting advantage, but when output may be at least as > high), cross-winds, narrowing of the course, or short anaerobic bursts among > riders at the front - all of which cause splits (i.e. PDR>1 at these > points). In a competitive situation, instabilities occur frequently > causing temporary splits at various places in the peloton, but these are > often closed when the cause of the instability has ceased. Sorting thus > occurs according to some combination of peloton rotations in which stronger > riders are able to get to the front and the continual splits in the peloton > at points of instability and reintegrations. I would need to develop the > model some more to show this as an equation (though I touch on a basic > version of it in my Appendix). > > For sperm, I! don't know what the initial state of the aggregates are when > they begin their travels, but I am assuming (perhaps quite incorrectly), > that there is some initial phase in which they are mixed (such as cyclists > on a starting line), and then they begin to sort as they increase speed. > During the process, they aggregate like cyclists because a broader range of > fitness levels can aggregate together (causing an effective narrowing of > fitness). As in a peloton, there are instabilities that allow for > continuous re-adjustments to the relative positions of all the sperm, and > over time they begin to sort into groups where each have fitness levels > closer to the average. This is my hypothesis, at least. > > On the second last question, there would be an advantage to sperm among the > first pulse aggregation if all the pulsed aggregations do not mix first, but &g! t; the principles apply to each aggregation. However, I don't! know wh ether > there is some other process of mixing first among all the pulses of sperm > aggregations before they begin traveling (I imagine I could find the answer > in the literature), in which case there could easily be a sperm in, say, > the second pulse, which could end up impregnating the egg. > > I don't know about the kamikaze sperm - I'll leave that one for now! But I > do remember that scene from the movie as clear as day! > > In any event, my aim is really to ask the question - are there energetic and > coupling principles that allow sperm to end up in groups which otherwise > appear to have occurred because genetically related sperm can somehow > identify each other? I am really only suggesting the existence of some > dynamics of the sperm aggregations that could be studied for, which don't > yet appear to have been addressed. > > Hugh >
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org
