Update on the Cytochrome C question

[bob b]

Some time ago on a thread devoted to the analysis of cytochrome C protein sequences I raised the point that many if not most protein sequences would be non-functional. Thus, it would be likely that there would be no feasible “pathway” to convert one cytochrome C sequence into another since any change to a sequence would be unlikely to result in a functional protein.

There now appears to be evidence that my intuitive “feel” for this issue may have been correct, “in spades”.

http://www.arn.org/docs/behe/mb_smu1992.htm

[rbisback]

"But the believer in the universal application of physical law is stuck. He must maintain, against the evidence, that different protein classes, like cytochromes and immunoglobulins, found their way by raw luck through the vast, dark sea of nonfunctional sequences to the tiny islands of function we observe experimentally. He must maintain, without any evidence, that Mesonychid gave birth over time to the whale. And why, we ask, must he maintain these positions against impossible odds and without supporting evidence? Because, he replies, I can only measure material phenomena, and therefore nothing else exists."

Interesting!

[shima]

>>We proceed in the following manner. If any of ten amino acids can appear in the first position of a given functional protein sequence then the odds are 1 in 2 that a nondirected search will place one of the allowed group there. If any of four amino acids can appear in the second position then the odds are 1 in 5 of finding one of that group, and the odds of finding the correct amino acids next to each other in the first two positions are one-half times one-fifth, which is one-tenth. Suppose in the third position there is an absolute requirement for G. Then the odds of getting a G at that position are one in twenty and the odds of getting the first three amino acids right are now up to one in two hundred. In this aspect it is like winning a trifecta in horse racing. Over the course of 100 amino acids in our small protein the odds quickly reach astronomical numbers.<<

Unfortunately, this research did NOT take into account changes at two or more positions SIMULTANEOUSLY. For example, the research might show that a G at position 3 is ABSOLUTELY NESSECARY for the protein to fold. Thus only x-x-G-x will fold. However, the research only shows that the amino acid G cannot be replaced with ALL OTHER AMINO ACIDS held constant. Suppose position 4 has an S amino acid. The reseach shows that x-x-T-S-x... doesn't fold. What it DOESN'T show is if x-x-T-G might fold. In other words, it only varies one amino acid at a time. Perhaps the G-S and the T-G combo might fold as well. Perhaps other combinations fold as well.

The research is extremely limited because they only tried LINEAR combinations (varying one varibale at a time) instead of the far more accurate but more time consuming non-linear 2nd, 3d and higher variations (ie variations in two, three or even more positions at a time). So the research only shows the change in a single variable. This represents the LOWEST POSSIBLE CHANCE in their calculations . Thus, the 1 in 10^65 is the LOWEST POSSIBLE CHANCE that a usable folding protein is formed by evolution. This chance will only INCREASE with the availability of 2 or more changes in position.

Now, for only the second order variation to be researched, the scientists will have to produce 20 (number of different amino acids) * 100 (number of possible locations for variation 1) * 20 (number of different amino acids) * 99 (possible locations for variation 2) = 3,960,000 possible combinations to investigate. This is 20*99 times as much as the original research.

[bob b]

Shima,

Why be so timid? Why not assume that all mutational changes occur as more than one simultaneous change at a time so that if it takes 12 changes to transform one sequence to another one that all 12 might occur simultaneously, thus avoiding any intermediate "prohibited" protein variants?

After all, given enough time anything is possible, right?

Seriously, the scheme you propose only makes the probabilities less likely. The reason has to do with combinatorial math, meaning that the number of possibile combinations taken two (or more) at a time is vastly greater than for one at a time. Since the number of functional sequences for proteins is a constant, the probability of hitting a new functional one is less for the two at a time case than it is for the one at a time case. The only virtue of the two at a time case is the remote possibility of "leapfrogging" a blockage, which is the equivalent of being able to drill through a blocked passageway when following the wrong pathway inside a maze.

The situation is even worse than described above, because we should more accurately be interested in not only transformation to a new functional protein but one that infers greater survival capability on the lifeform. The reason is that the new protein must be able to be spread throughout the population for it to be effective as an evolutionary mechanism. If it confers a reduced survival or reproductive value on the lifeform then the odds of it spreading throughout the population are drastically reduced (or as one wit put it, "the chances are somewhere between zero and none".)

I did not expect any believer in evolution to be fazed by the enormous unlikelihood of molecular transformation of protein sequences, but I was interested in hearing some of the subjective arguments that would be advanced in order to attempt to explain away the obvious implications of the protein experiments cited in the article link that I posted.

[rbisback]

Hmm, seems intelligence is required to explain the process, why is it such a reach to believe intelligence MADE the process in the first place?

[Mr. Ben]

The problem with this argument is two fold, it's the old backwards odds calculation, as well as ignoring the stepwise processes of evolution.

With the first problem, we have the old "what is the probability that rbisback or bob would be born" argument. It is highly improbable that Bob's and RBisback's parents would live to maturity, meet, decide to mate, and happen to have exactly the child that grew up to be who we now know as Bob or RBisback.

However, it is NOT improbable that two people amongst billions would meet, and that these two people would have A child, and that it would be male and that somebody similar to Bob and RBisback might be born. That is precisely what did happen.

The fact is that there are thousands of proteins in your body, and all of them suffer mutation. What behe's doing in this ridiculous article is calculating the odds of a particular gene having the sequence it does. Though the ratio of functional to non-functional proteins is high, there is no reason why we should expect that any given functional protein must have the precise sequence we see now. We can postulate a huge set of functional proteins, of which the ones we do see are only one of many, just as we know of a huge set of potential offspring of any human couple.. of which RBisback and Bob are two ordinary highly likely examples.

Not only this, but we are treated to the usual creationist sidestepping of the issue of process. Just like we can't tell how unlikely somebody like Bob or RBisback occuring is unless we know about how likely it is for any particular type of person to be born, we also must know that people having children is likely, that people meeting, falling in love, and marrying is likely. If these processes and events are likely, then that makes the occurence of the birth of new individuals of which Bob and RBisback are an average and expected subset quite likely.

Likewise, if we know that stepwise mutation, inversion, or crossing followed by selection for fitness is a common occurence in coding genes, that means that the stepwise (and even the sudden crossing) of new coding genes building folding protiens is pretty darn likely. It doesn't tell us that any 'particular' protien is likely, only that a set of proteins constrained by natural selection might be the result.

So there we have it, we must know how many possible alternatives to any particular protien that "could have worked" there are to get any sort of estimate as to the probability of a particular sequence... and we must take into account the actual process by which these protiens are sequenced (mutations of various types plus selection).

Without these, any estimate of probability is nonsense.

[bob b]

Ben,
It is interesting that you would attack the conclusions of evolutionary scientists who specialize in protein analysis. Behe cites numerous studies in the link I provided. Your comments seem to imply that it is Behe alone who has done the calculations.

Having said that I do agree with some of your points. The studies do not provide conclusive "proof" that random mutation does not play a role in generating new protein sequences. However, at a minimum the studies illustrate that the random mutation plus natural selection hypothetical mechanism is primarily a subjective rather than a quantitative argument. In other words it may sound good to some, particularly if one is committed to the belief in major transformations being a reality, for if not this mechanism then what?

For me the value of the studies is to confirm my initial subjective feeling that molecular analysis was ignoring the central thesis of evolution that natural selection acts on "improvements" to spread them throughout the population. When a phylogenetic analysis is performed on a molecule such as cytochrome C (my favorite) one is tacitly assuming that a transformation has taken place, which means that any intermediate steps required to effect the molecular transformation must have been successful and were due to a series of steps, each having greater evolutionary success than the previous one.

The studies cited demonstrate what a large pill this is to swallow. They do not constitute conclusive "proof" to be sure, but for people who are not absolutely committed to the "fact" of major transformations, surely they provide a wakeup call that perhaps major transformations may not be as inevitable as they previously thought.

More will have to be done before the issue is finally settled, and this may take decades considering the difficulties involved, but I can't help wondering whether these findings are analogous to the previously mentioned example of "combustion theory" where the option of "negative mass" was seriously entertained by some of the biggest names in the field to their eternal ignominy.

[Stratnerd]

Bob B.,

Rather than go over all the tangential material provided here, please, since you've worked on it so hard, give us a the steps in terms of amino acid changes going from organism X to organism Y that are physiologically impossible. This is the simplest way to cut to the chase, eh? Let's go from chimp to human for starters and show us how those transformations in cyt c are impossible.