In 1966, there was a conference held at the Wistar Institute in Philadelphia and this problem was first explicitly formulated. There was an MIT scientist, a professor of engineering, Murray Eden, a computer engineer who raised the problem this way. He said that, “No currently existing formal language can tolerate random changes in the symbol sequences which express its sentences. Meaning is almost invariably destroyed.” Now this conference followed a picnic in 1965 on the lawn of a famous MIT physicist named Victor Weisskopf. And at the picnic an argument broke out between the physicists, the engineers, the computer scientists, and the mathematicians on one side and the biologists on the other. The biologists have been chatting up all the exciting things that had been going on in their field. 1965, that’s eight years after the sequence hypothesis. By this time the whole gene expression system, the system of protein synthesis, had been pretty well elucidated. The biologists now knew that the information in DNA was directing the construction of proteins. And the physicists and engineers at this conference were saying, “That’s very cool, but if that’s the case we don’t get how your neo-Darwinian Theory really works. If DNA is a section of digital code like we use in computer science, how do you randomly change the sequences and hope to get something functional on the other side? We don’t see how that would work.” This is Murray Eden’s point. You’ll degrade. Meaning is almost invariably destroyed if you have lots of random changes. Biologists said, “We see your point, but it could be that there’s lots and lots of ways to go right. Lots of ways of organizing the A’s, C’s, G’s, and T’s that will result in functional proteins. That functional proteins are common among the space of possible amino acid combinations.” Is this sounding familiar? It’s the opposite of what Doug Axe ended up proving. This was the discussion they ended up having at the Wistar Institute. And here’s another way of understanding the problem. In the English language, if you have a 12 letter sequence, for every one word that’s meaningful or sequence of letters that’s meaningful, there are 100 trillion possible ways of arranging the 26 letters of the English alphabet that will not be meaningful, that will be gibberish. The ratio of meaning to gibberish is extremely low. And what the biologists were hoping for was that, in the biological case, the ratio of function to gibberish would be high. So that if you had random changes, you would almost inevitably stumble on to some new thing that was also functional. You skipped from one island of function to another fairly easily. In 1966, they had this debate and it all came down to the question of rarity of genes and proteins in sequence space. The computer scientists and the mathematicians said, “We’re pretty skeptical. If it’s a true linguistic system, if it’s an information-carrying molecule, probably meaning is gonna be rare. Function is gonna be rare.” But the biologist said, “Well we don’t know that. It could be common.” And nobody knew. And that’s where Doug entered this fray about 23 years later. There were more indications of rarity, but nobody had definitively pinned it down. He mentioned a lab in, at MIT, Robert Sauer, who first started to do these site-directed, mutagenesis experiments that Doug later did and perfected. And he was getting measures of extreme rarity, but not even as much as what Doug later determined.
So I’d like to illustrate the problem this way. It’s called the problem of combinatorials. You’ve got four dials and ten digits on each, you get 10,000 possibilities, only one combination. If you’re are a thief out there trying to crack this lock, maybe it’s on a bike outside of Mote Hall, is it more likely that you’ll, that you’ll crack the lock by chance or fail? It’s a trick question isn’t it? What else do you have to know? You have to know how much time? How many opportunities? If you have enough opportunities to do 5001 possible combinations it becomes more likely that you’ll succeed than fail. In which case the chance hypothesis becomes more likely to be true than false. But what if the thief encounters this lock? Now it gets a lot harder doesn’t it? I did some math on this and found that if the thief lived for a hundred years, changed one dial every ten seconds, took no potty breaks, had no dates, never ate dinner, or anything else, he could sample about 3% of the total 10 billion possible combinations in a hundred years. In that case it’s much more likely that the thief will fail than succeed in randomly searching for the combination in which case that chance hypothesis is more likely to be false and true. Doug posed this question: how common or rare are the functional sequences of proteins among all the possible combinations? One in 1077. That’s like a lock with 10 digits on each dial and 77 dials and you’re looking for one possible combination. But in the biological context what do you have to know? It could be that you have lots of opportunities to find that functional protein or a functional fold. There was another question that had to be asked. It’s effectively the question “How much time or how many opportunities?? And it is how many organisms have ever existed on planet earth in the history of life? Because you only get a randomizing of a section of genetic code when there’s a replication event. That’s an opportunity for a mutation to take root. So if you know how many organisms have lived, then you know the maximum number of opportunities you would have to randomly change any given section of the genetic text. So that’s actually something fairly easy to estimate. Biologists have done that. And the answer is there are about 1040 organisms in the history of life. Now this is a slightly different set of numbers than in the origin-of-life case because now we’re talking about biological evolution. But 1040 organisms have lived. That’s a lot of opportunities to mutate something. But is 1040 a big number compared to 1077? No. That’s an incredibly small number. What happens when you divide exponents like that? You subtract. So what’s the ultimate ratio, in other words, how much of the combinatorial space of possible arrangements can be searched? One over 1037. That’s 1/10 trillion trillion trillionth. We’re gonna change the metaphor to a great big haystack. A little tiny needle. We’re looking for it somewhere in that big haystack. But we can only search the tiniest little smidgen of the haystack. 1/10 trillion trillion trillionth of it. In that case, is our random search more likely to be successful or a failure? Overwhelmingly more likely to fail, in which case again the chance hypothesis that that’s how it happened, that that’s how information was generated, is more likely to be false than true. Same kind of reasoning as before, just a little bit different math in the biological setting. But this is very tight, very rigorous. And of course you don’t just need one protein to build a new type of cell type, it’s much harder than this. So this was Doug’s result applied to biological evolution.
Dembski developed that design detection methodology, published it in the book with Cambridge University Press, 1998. Back to the ancient Greeks, to Cicero in the Roman philosophy, the medieval philosophers, Aquinas and others. Newton. All through the history of Western thought, scientists and philosophers have tried to crack this problem. How can we tell when something is the product of an intelligence as opposed to undirected natural processes? Dembski formalized a method that allowed us to distinguish between the two. I think it’s a huge intellectual achievement for which he has received only brick, bats, and opprobrium, and condemnation, and ad hominem abuse. But someday Bill Dembski will get his due. It’s an incredible discovery, an incredible innovation in mathematical reasoning. What Doug has done is something of equal magnitude, because this is the key question. The plausibility of the Darwinian mechanism relies on a random search method before natural selection can kick in. Natural selection can only select what random variations or mutations have first produced. The innovation in the mutation selection mechanism was always thought to come from the variations, the mutations. Natural selection is just a mechanism of non-random death. If the variations are favorable to survival they exist and move on. If they’re not, the organism dies out. Innovation comes from the variation and the variations must come from the mutations. So the key question as to the plausibility of that mechanism was the question of the rarity of the functional genes and proteins among what we call combinatorial sequence space, all the other possible ways of arranging the constituent parts in these big molecules. And Doug Axe cracked that problem. He got the number. He did a very rigorous method.
And interestingly an Israeli scientist named, Daniel Tawfik, who was also in the lab in Cambridge – same lab where Doug was for time – very secular guy, no friend of intelligent design, has been publishing papers based on his own research with proteins that have not only been confirming Doug’s result but they have been extending it. They have been showing that it’s perfectly generalizable. Doug studied one crucial protein. Tawfik has studied an ensemble of about 20 different globular proteins. And what he’s done is he’s induced mutations into the DNA sequences that get transferred into the protein structure, and he’s found that invariably after between 3 and 15 mutations – it varies a little bit from protein to protein – the structure, the structural stability of those protein folds is lost and unwound. They become thermodynamically unstable. And yet we know that it would take many more mutations than just 15 in the best case to build a new protein fold. So Doug’s conclusion that the protein folds are too rare and too isolated in combinatorial sequence space to be reached by random mutation has been confirmed by another scientist taking a slightly different approach. And Tawfik’s work shows that Doug’s results can be generalized not just for one protein but for whole classes of proteins. Tawfik, a completely secular guy, hates intelligent design proponents, suspected Doug of being one, characterizes the origin of novel protein folds as quote, “something close to a miracle.” We have no evolutionary explanation of the origin of the smallest unit of biological innovation: the protein fold. And if you can’t explain that, good luck explaining human behavior, sexual behavior, or religious belief, or all the things that evolutionary psychology is trying to explain, let alone body plans or new organs and tissues. If you can’t explain the protein fold you haven’t gotten anywhere.