Editor’s note: We are pleased to welcome a contribution from Dr. Sanford, who is Courtesy Associate Professor, School of Integrative Plant Science, Cornell University.
My colleagues and I recently published a paper in Theoretical Biology and Medical Modeling, “The Waiting Time Problem in a Model Hominin Population.” It is one of the journal’s “highly accessed” articles. A pre-human hominin population of roughly 10,000 individuals is thought to have evolved into modern man, during a period of less than six million years. This would have required the establishment of a great deal of new biological information. That means, minimally, millions of specific beneficial mutations, and a large number of specific beneficial sets of mutations, selectively fixed in this very short period of time. We show that there is simply not enough time for this type of evolution to have occurred in the population from which we supposedly arose.
Historically, Darwin-defenders have argued that time is on their side. They have claimed that given enough time, any evolutionary scenario is feasible. They have consistently argued that given millions of years, very large amounts of new biologically meaningful information can arise by the Darwinian process of mutation/selection. However, careful analysis of what is required to establish even a single genetic “word” (a short functional string of genetic letters) within a hominin genome shows just the opposite. Even given tens of millions of years, there is not enough time to generate the genetic equivalent of the simplest “word” (two or more nucleotides). Even in a hundred billion years, much longer than the age of the universe, there is not enough time to establish the genetic equivalent of a very simple “sentence” (ten or more nucleotides). This problem is so fundamental that it justifies a complete re-assessment of the basic Darwinian mechanism.
In my book Genetic Entropy, I have previously outlined the waiting time problem (for example, see the 2014 edition, Chapter 9, pp. 133-136). My calculations there, and calculations published by others (Behe, Snoke, Axe, Gauger et al.), all demonstrate the same basic problem. (For a complete literature review, see the link to our new paper given above.) What this new paper provides is an independent validation, by a totally different method, of the previous works done by Behe, others, and myself.
In our paper we examine the waiting time problem in a new way, employing state-of-the-art, comprehensive, numerical simulations to empirically document the time required to create a specific string of mutations. This method is an alternative to employing mathematical approximations, and is preferable for various reasons outlined in the paper. Our empirical experiments realistically enacted the establishment of short genetic sequences within biologically realistic virtual hominin populations. These experiments demonstrate the limits of the classic neo-Darwinian mechanism in a clearer, and more compelling way. Of special significance, we show that as genetic “word size” increases linearly, waiting time increases exponentially (see Table 2 in the new publication).
The waiting time problem has four basic elements. First, in a small population it takes a very long time for any specific nucleotide (genetic letter) to mutate into a specific alternate nucleotide. Second, it takes vastly more time for a given string of nucleotides to mutate into a specific alternative string of nucleotides (as is required to create a new beneficial genetic “word”). Third, any specific new word that arises is quickly lost due to genetic drift, and so must arise many times before it “catches hold” within the population. And fourth, even when the new word catches hold, it takes additional time for natural selection to amplify the new beneficial mutation to the point of fixation within the population.
Our paper shows that the waiting time problem cannot honestly be ignored. Even given best-case scenarios, using parameter settings that are grossly overgenerous (for example, rewarding a given string by increasing total fitness 10 percent), waiting times are consistently prohibitive. This is even for the shortest possible words. Establishment of just a two-letter word (two specific mutations within a hominin population of ten thousand) requires at least 84 million years. A three-letter word requires at least 376 million years. A six-letter word requires over 4 billion years. An eight-letter word requires over 18 billion years (again, see Table 2 in the paper). The waiting time problem is so profound that even given the most generous feasible timeframes, evolution fails. The mutation/selection process completely fails to reproducibly and systematically create meaningful strings of genetic letters in a pre-human population.
Other authors have published on the waiting time problem and they have consistently acknowledged its reality, but some have then tried to minimize the problem. In those cases, the authors have first shown the waiting problem is serious, but then go on to invoke very special atypical conditions, seeking to reduce waiting times as much as possible. This is evidently in the hope of saving neo-Darwinian theory. But when these “special conditions” are carefully examined, in every case they are far-fetched and ad hoc.
When the dismissive authors use the same formulation of the problem as we used in our paper, they see the same prohibitive waiting times (see our paper’s discussion). For example Durrett and Schmidt (2007) model a human population of 10,000, just as we do. They show that for a specific set of eight required mutations (which must arise in the context of a specific genomic location), the waiting time is 650 million years. But most readers will miss the fact that this is just their estimated time to the “first instance” of the string. Elsewhere in their paper they acknowledge that the establishment and fixation of the specific set of mutations would take 100 times longer than the first instance (when they assume a 1 percent fitness reward). This would be 65 billion years! Using the same parameter settings (and applying a 1 percent fitness reward) our own experiments give waiting times of the same magnitude. Likewise, when Lynch and Abegg (2010) specify a population of 10,000, and when two specific mutations are required, they get waiting times exceeding 10 million generations (see their Figure 1). Assuming twenty years per generation for a human population, this is more than 200 million years (see our paper’s discussion).
What will the primary counterargument be to the waiting time problem? The primary objection is, and will continue to be, as follows. Within a small population, a given string of letters cannot arise in a specific location without a prohibitive waiting time, yet somewhere else in the genome good things might still be happening. For example, if one is waiting for the sequence ATCG to be fixed in a specific genomic location, it will require very deep time, but it will take no time at all if one is waiting for ATCG to arise anywhere in the genome. Indeed, many copies of ATCG are already in the genome. This argument has three problems.
First, it ignores context. The sequence ATCG by itself is not useful information. It can never be beneficial (and hence selectable), except in a very specific context. Consider randomly changing one word in an encyclopedia — will it consistently improve the text, regardless of where the change is made? All information is context-dependent. For example, if you have an executable computer program, inserting a certain random string of binary digits could conceivably improve the program’s information content. But in such a very unlikely case, it would only be beneficial within an extremely specific context (location). When inserted out of context, the same string would almost certainly be deleterious.
Second, when we broaden our view to include the whole genome, we have to consider the problem of net loss of information, due to a multitude of nearly neutral deleterious mutations that are happening throughout the genome. Random mutation will cause ubiquitous genetic damage (especially in deep time), which will greatly overshadow the few rare strings that might arise in just the right context and might be sufficiently beneficial to be selectable.
Third, invoking “good things that might be happening in other parts of the genome” is essentially sleight of hand. Other potentially beneficial sets of mutations in other parts of the genome will each have their own waiting time problem. This is not a reasonable explanation for the origin of the massive amount of integrated biological information that is required to change an ape into a man (i.e., millions of complementary nucleotide substitutions established and fixed within the source hominin genome, in very little time).
Given that higher genomes must continuously accumulate deleterious mutations (as I show in Genetic Entropy), and given that beneficial mutations are very rare (as shown by the famous Lenski LTEE project, and also as shown in Genetic Entropy), and given that evolution cannot create meaningful genetic words (even given deep time), it seems that neo-Darwinian theory is coming undone on every level.