1. At first glance, this appears to be a strong argument for evolution. Indeed, I found it troubling for a while.
2. However, we are still learning, and it is hard to know when a part of the DNA is beneficial or harmful. How do we know for sure what is an error? Even the so-called junk (nonfunctional) DNA may have a function that we do not yet understand. For example, an article in Science (4 April 1997, page 39) suggests that "satellite DNA," which some researchers regard as nonfunctional, may have a function. See also Science 1994, Feb 4 pp. 608-610. This whole area is sufficiently new that it is best to wait a little longer. We should also look at shared properties of the DNA between many species, and see if they can consistently be placed in an evolutionary tree. If not, then this calls into question the human-ape connection.
3. Mutations are not completely random. It's possible that the same kinds of mutations tend to occur in the same way (for example, where the DNA folds, or whatever. Dan Hughes also suggests that DNA might tend to adopt a low energy state.) This could explain many common errors. Note that mutations in a population can be expected to obey more regularities than those among individuals, because of the similarities in survival benefits and the laws of large numbers.
4. Maybe the Lord inserted those similarities for a reason we do not understand. They could even have been inserted as tests of our faith. The Lord does not force any to believe, but gives opportunity to doubt for those who are seeking it.
5. Another possibility is that the Lord, when he cursed Adam and Eve after the fall, also cursed all life by introducing errors into the DNA. One could expect that similar species were cursed in a similar way, out of fairness.
6. Edward Max's argument is based on the fact that these shared sequences are really errors, that is, mistakes. It seems strange to call something an "error" when it occurs in a nonfunctional part of the DNA. Since that part of the DNA is nonfunctional, it doesn't matter what occurs there, so there is no justification for calling some of the sequences errors.
7. Something has to appear in the nonfunctional part of the DNA. Why should it be one thing rather than another? Just by chance (accident) there are likely to be sequences that resemble genes, but this says nothing about their origin. Do we expect that the Lord would have deliberately avoided common subsequences at common locations just so that we would not think there was common descent?
8. Let us consider humans and apes. Since they are so similar, one would expect that they had many similar genes at the creation in similar locations in the genetic material. Also, with the change in environment since then, one would expect that some of these genes would no longer be necessary, such as the gene for synthesizing vitamin C, and that there would be a few such genes in common between apes and men. Now, point mutations arise all the time, and if they are fatal or harmful, they will disappear from the population (have a small frequency). If they are neutral, they can be passed on. So it is reasonable that point mutations inactivating the vitamin C synthesizing genes would occur in both humans and apes, and be preserved in both, since these genes have little benefit now. Thus we would get a pseudogene in the same location in humans and apes. It could have been present in the individuals on the ark, for example. This probably would occur for a few other genes, as well. For organisms that are less similar, this still could happen, but less often. So we would expect to find a pattern of common pseudogenes that reflects the similarity between organisms, but not as an evidence of common descent.
Actually, the LGGLO pseudogene (an inactivated Vitamin C synthesis gene) has been found in one human so far and no apes, according to Edward Max, but in his essay he predicts that it should be found in apes, too. In fact, given the similarity in the DNA of humans and apes, that is a reasonable prediction. There are other examples of common pseudogenes that he says have been found in humans and apes, but I do not know yet if they occur in exactly the same form in humans and apes, or in how many individuals they have been found, or how many base pairs they have. Even if they do occur in the same form, this would be a problem for evolutionary theory, I think, because neutral mutations tend to be eliminated from populations, according to the talk.origins archive, and one would not expect a neutral mutation to persist for so long. Some neutral mutations can spread to the whole population, but this generally takes a very long time and has a low probability. (The chance that this will happen is proportional to the frequency of a neutral mutation in the population.) By seeing how much variation exists between copies of the mutated region in different organisms, one could estimate its time of origin and in this way check evolutionary timetables, since non-functional DNA probably mutates at constant rates. One would not expect to find a large number of shared neutral mutations among all humans and apes, in any event. Also, one would expect to find neutral mutations that had only spread to half or a fourth of the population, too, since they spread so slowly when they do spread. Furthermore, one would expect a gene to be inactivated in many different ways, so that exactly the same form should not be found today in all individuals. This would seem to imply a very severe population bottleneck at some time in the past.
9. It is even possible that the lack of ability to synthesize vitamin C could be an advantage in certain situations, although this appears unlikely. It could be that individuals without this ability would be forced to move to a different location, and this new location could turn out to be a more favorable habitat. The same could be true for the loss of some other genes.
10. Before being created, life was an idea in the mind of God. The relationship between the created life forms reveals something about how ideas develop in the Divine mind. We cannot say in advance how these ideas develop, which makes it difficult to draw conclusions about how the various life forms were interrelated at the creation.
By the way, concerning the junk DNA, it should also be mentioned that this DNA helps to concentrate certain kinds of mutations (crossovers, recombinations) at certain places in a protein molecule, and this can be a valuable function. The recombinations are likely to occur between the pieces of a gene, and not in the midst of it. This could have a useful function. However, this should only be dependent on the length of the junk sequence, and not its content.
The talk.origins site about this subject is carefully reasoned and does a fairly good job of presenting the creationist side, by the way, although it becomes somewhat biased and sarcastic at the end, in my opinion.
I wanted to comment in more depth about the effects of the curse. One of the purposes of the Creation is to illustrate spiritual truth. Before the entrance of sin, a perfect creation could faithfully represent spiritual reality. After the fall of man, this was no longer so. Now it would be necessary for the plants to bear thorns and thistles to teach spiritual lessons, and for the soil to be difficult to work. Jesus in his parables often referred to the things of nature as illustrations of the workings of good and evil. It is sad that the innocent animals had to bear the effects of sin which they had not caused, but their sufferings do help to bring to the mind of man the terrible effects of evil and lead him back to God.
The key observation is that according to population genetics, if a neutral mutation spreads to a population of size N, it takes an average of 4N generations to do it. So if we see a number of neutral mutations spread to the whole population in M generations, this suggests that the average population size since the neutral mutations was at most about M/4.
Now, let's consider the time since the Cambrian explosion about 570 million years ago, and consider the assumed evolutionary ancestors of man during this time. This is about 5 * 10 8 years. We can estimate an average generation time of at least half a year; even if organisms can reproduce faster, a typical individual will be born after more than one reproductive cycle. This gives about 10 9 generations. Suppose that we find shared pseudogenes going back all the way to the beginning of the Cambrian explosion but not much farther. These would have spread to the whole population in 109 generations, implying an average population size of about 2.5 * 108. Now, we can also assume that shared pseudogenes will be found from later periods, which would have had to spread correspondingly faster. This implies that the average population size was decreasing with time, so we can say that about 1.25 * 108 is a reasonable value, which we round to 108. Humans have about 3 * 109 base pairs in the genetic material, which has been increasing in length. Thus its average length would probably be less than 2 * 109. Thus there would be about 2 * 1017 base pairs per generation. The rate of mutation is estimated to be one in 1010 to one in 1012 base pairs per generation. Let's use one in 1011. This means 2*106 mutations per generation.
We distinguish between major and minor mutations, as explained in A Theory of Small Evolution.. A major mutation is one that changes the shape of a protein. This would include increasing its length. Now, of 2*106 mutations per generation, maybe 9/10 would occur in the nonfunctional region, leaving 2*105 in the functional region per generation. Of these probably half would be minor mutations or have no effect at all and the other half would change the shape of a coded protein, giving 105 major mutations per generation.
Now, we estimate the probability that a protein with a changed shape will have a beneficial function in the organism. In reality this appears to be vanishingly small, but we will be very generous. In the immune system, it generally takes about 100,000 antibodies before one is found that binds to an invading organism (antigen). This is a highly specialized system, and a protein generally has to do more than just bind to something in order to be beneficial, so we will say one in 106 major mutations is beneficial. From 105 major mutations in all, this implies that about 1/10 per generation are beneficial. Even most beneficial mutations do not fix (reach a frequency near one) in the population, so we will assume that 1/100 of these fix, leading to 1/1000 per generation that fix. In 109 generations, this leads to 106 beneficial major mutations that fix.
Actually, the one in 106 figure is much too high because antibodies all have the same shape, roughly speaking. Proteins have many different shapes, and these also must match precisely for the proteins to interact. It takes a lot of information to specify a three-dimensional shape, so we can say that the probability is one in 106 that the shape will be right for an interaction. The protein also has to be able to "unstick" from its partner at the right time; this requires a carefully constructed system, and we will say another probability of one in 106 for this. This leads to a probability of one in 1018 so far to interact with one other protein. Generally a protein has to interact with at least 2 others, so we can square this to obtain one in 1036. Now, even if a protein interacts, in all probability this will be a harmful interaction, so we can say that the chance of a beneficial interaction is much less than one in a thousand, leading to a probability of one in 1040 overall. This agrees with the fact that the smallest genes typically have about 150 base pairs, and 4150 is about 1090. If this (admittedly approximate) one in 1040 figure is anywhere correct, then the probability of a single beneficial major mutation in the history of human evolution is essentially nil.
Let us consider in more detail the mutations that add a base pair to a gene. If this is added in the middle, it will probably change the shape of the gene significantly; especially will several such mutations in the middle have this effect. Now, a mutation at the end might just tack on a new amino acid to the polypeptide sequence but not change its shape. Each gene is composed of pieces, an average of three pieces, each typically in length from 50 to 1000 amino acids, that is, 150 to 3000 base pairs. Let's use 1000 as a typical value. An addition of a base pair in the middle of a piece will mess up the 3-base pair codons terribly and produce a completely new protein with a new shape. The chance that a mutation will occur at the end of this sequence is about one in 1000, among those mutations that add a base pair. Three such mutations will add an amino acid. I don't know what fewer than three will do, but they (and the third one) could be neutral, harmful, or beneficial. The first two kinds (neutral and harmful) will tend to be eliminated and not fix in the population. Even if the mutation is beneficial, after a few amino acids are added on the end, the shape of the protein will probably change due to forces involved in protein folding. Thus we will need about as many changes of shape as their are amino acids added (to within a small constant factor). If the shape does not change, then the function will probably not change much, either, which would mean that not much change in the organism would occur.
When we have a protein with a new shape, the old function will no longer be performed. This will probably be harmful to the organism. So in order to allow this, we will have to first copy the gene by a mutation. This copy will be neutral and probably eliminated, too. If not, it will be beneficial, so when the shape changes, the mutation is probably harmful and will be eliminated. Even if not, we will end up generating a sequence of genes, each differing by only a small length from the preceding members. This whole scenario is becoming ridiculous.
Another possibility is to imagine that pieces of existing genes got shuffled around to produce new genes with beneficial functions in the cell. A problem with this is that it by definition involves huge changes in protein shape, not compatible with evolution's gradual change philosophy. Also, such large changes in structure are very unlikely to produce anything useful. Finally, such shuffling of genetic material is likely to be fatal, so it can't happen very often.
Anyway, how many of these length-increasing mutations can be beneficial? Note that only about 1 in 1000 of mutations can add an amino acid on the end. Many mutations will be point mutations, and some will delete an amino acid. Even of those that add an amino acid, some will be neutral and some harmful. So if we assume that 1 in 1000 of the length increasing mutations at the end of a gene are beneficial, it seems a plausible estimate and leads to the same rate of evolution as mentioned earlier. If anyone has better figures, I would be happy to have them.
The question is now whether this is an adequate number. For this we will use the one in 106 probability estimate that a major mutation will be beneficial; of course, this is undoubtedly much too generous, and one in 1040 or less is more likely. The human genome has 3*109 base pairs, and maybe 3*108 are functional. To create this amount of genetic material, we have to increase the length of a gene about this many times (to within a factor of 2 at least), and each such increase would involve a change in shape of the coded protein. This can be avoided by copying genes and mutating the copies, but I don't think that this will reduce the number by very much. Now, minor mutations (those that do not change the shape of the protein) probably don't have much effect on the organism anyway, so we would expect that most of the change to the genome was due to major mutations. Thus we only have 106 beneficial major mutations to account for 3*108 functional base pairs (or maybe half this number). We see that this number is entirely inadequate.
To get around this, we need to increase the population by a factor of at least 102. This would mean that we probably would not see many (probably not any) shared pseudogenes or other shared neutral mutations between different species, depending on how the population size varied with time. Furthermore, there should at least be some species (according to accepted evolutionary theory) that have persisted relatively unchanged for hundreds of millions of years with huge populations; these should have few, if any, shared pseudogenes and considerable variation in their non-functional DNA.
We now give justification for the statement that the non-functional DNA should be randomized by mutations, since this is not obvious. For the sake of illustration, suppose that there is one point mutation per base pair for every 200 million years. This figure is obtained from the following quotation from Introduction to Evolutionary Biology from the talk.origins archive:
Li and Graur, in their molecular evolution text, give the rates of evolution for silent vs. replacement rates. The rates were estimated from sequence comparisons of 30 genes from humans and rodents, which diverged about 80 million years ago. Silent sites evolved at an average rate of 4.61 nucleotide substitution per 109 years. Replacement sites evolved much slower at an average rate of 0.85 nucleotide substitutions per 109 years.Non-functional DNA behaves much like silent sites with respect to mutations. A rate of 4.61 substitutions per billion years means about one per 200 million years. Suppose we consider half this many years, so about half of the base pairs will have mutated during this time. This would be at most 100 million generations, probably. Let's also suppose that the population is large enough (significantly larger than 25 million) so that probably none of these mutations will have spread very far relative to the whole population. Let's assume that at the beginning, all the non-functional DNA in the population is identical. We compare two random individuals after 100 million years. Each will have about half of the non-functional DNA mutated in this time period (considering only point mutations), actually somewhat less. Thus one-fourth of the base pairs will be mutated in neither individual, one-fourth in both individuals, and one-half in one but not the other. The first one-fourth of base pairs will agree between the two. The next one-fourth will probably disagree about half of the time if each mutation spreads to only half the population on the average. The other half will disagree. So we get about 9/16 disagreement in the non-functional DNA. As time passes, this figure will get larger and approach 3/4 since their initial DNA will differ as well.
According to population genetics, if a mutation does spread to the whole population, it takes on the average about 4N generations to do so, where the population size is N. However, due to the existence of partially isolated sub-populations, in reality this should take much longer than 4N generations. In order to get a significant amount of disagreement, we argued above that the population must be significantly larger than 25 million, since point mutations that do spread to the whole population will probably then take significantly longer than 100 million generations to do it. However, in fact a population significantly under 25 million should still suffice for this argument. This is a reasonable population size. For generation times of say 10 per year, the population would have to be 10 times larger.
Now, suppose enough time passes for 1/10 of the DNA to be subject to a point mutation. This would be 20 million years, or about 20 million generations. Then we would expect to get about 2/10 disagreement in non-functional DNA, even if the population started off with identical DNA. For this, the population would have to be somewhat larger than 5 million. (In fact, a much smaller population size would probably suffice.) This is still a much larger disagreement than is observed for humans. Since the generation time would probably be at least 10 years, humans would only need a population of about a million or more to reach this much disagreement in 20 million years. This seems to be an evolutionary puzzle, because the observed difference between humans seems to be about 1 in 200.
We could assume that population bottlenecks occurred during which the genetic material became uniform. But if this happened, the functional DNA would also become uniform in the population, except for beneficial traits that were heterozygous. It is my impression that such homogeniety is not observed as a rule today. It is difficult to recover genetic variety in the functional DNA, since it is tremendously expensive evolutionarily to construct alleles coding for proteins with new shapes. This is evidence that such extended bottlenecks did not occur.
The same computations apply to "silent" positions in the coding part of the DNA. A silent position is a base pair that can be changed without changing the amino acid coded by the gene. Since such positions have little or no effect on the organism, they behave much like non-functional DNA. Suppose a gene has k silent positions, and suppose the population is large enough so that no mutation is likely to spread to the whole population in 100 million years. Then in 100 million years, we should expect about half of these positions to have mutated. Assuming the population is large enough, each silent position will have about half of its DNA without a mutation and half of the DNA with a mutation after 100 million years. I estimate that at least 1/6 of the base pairs in a typical gene would be silent positions, and maybe as many as 1/3. Thus we should expect at least a 1/12 disagreement in the base pairs between alleles of two random individuals in the population after 100 million years, not even counting mutations that change an amino acid. A smaller average disagreement would indicate a recent origin of life or a severe extended population bottleneck, but the latter is unlikely to have occurred in all species in the traditional evolutionary scenario. The observed value of one difference in 200 base pairs among humans is significantly under the 1/12 value.
The actual amount of difference between humans is less than one in 200 base pairs, as the following quotation shows:
This quotation is taken from "Mapping Heredity: Using Probabilistic Models and Algorithms to Map Genes and Genomes (Part I)," by Eric S. Lander, Notices of the AMS July 1995. In fact, the average difference appears to be less than this:"Two genomes chosen from the human population are about 99.8 percent identical, affirming our common heritage as a species. But the 0.2 percent variation translates into some six million sequence differences."
This is taken from Chapter 3. DNA Typing: Statistical Basis for Interpretation. However, we will use the six million figure to be conservative. In order to get this amount of difference, we only need one mutation per thousand base pairs since the origin of the human race. Using the above rate of mutation, and recalling that most of the DNA is thought to be non-functional, it should only require about 200,000 years for this amount of difference to arise. This gives an age estimate of 200,000 years for the human race. However, the figure really must be significantly smaller than this, because there was undoubtedly some variation at the start, and because the actual rate of mutation is arguably faster than one per 200 million years. In addition, some mutations alter many base pairs at the same time.Can DNA typing uniquely identify the source of a sample? Because any two human genomes differ at about 3 million sites, no two persons (barring identical twins) have the same DNA sequence. Unique identification with DNA typing is therefore possible provided that enough sites of variation are examined.
We now estimate how large the human population must be in order for this 200,000 year estimate to be valid. A population geneticist sent me the following information about the propogation of neutral mutations:
There is a roughly 1/N chance of a new neutral mutation being fixed. There is a 2/N chance of it getting half way, and it will do so in roughly half the time (4N/2). Half of these will get the whole way, and they will take another 4N/2 generations. Better than that requires a more detailed calculation.Now, in 200,000 years there would be less than 20,000 generations. A neutral mutation will spread to the whole population in an average of 20,000 generations if the population size [N] is about 5,000. But if the population size is about 10,000 or larger, neutral mutations will on the average spread to only half of the population, which is sufficient for our age estimage to be valid. The value 10,000 for the human population is quite small, and it is reasonable to assume that the human population was generally much larger than this, implying that our young age estimate is correct. We suspect that a similar young age estimate is valid for other species as well, which I find difficult to reconcile with the accepted view of earth's history. For example, dogs and wolves also seem to have about the same or less genetic diversity as compared to humans.
There is another mechanism that can reduce this genetic variation, and this is the following: When a beneficial mutation spreads to the whole population, it will tend to carry along nearby base pairs, thereby reducing genetic diversity in its neighborhood. The size of the neighborhood is related to how fast it spreads. If it spreads to the whole population in about 1000 generations, then the size of this neighborhood is about 100,000 base pairs, since crossovers generally occur once in 100 million base pairs and there would be a thousandfold multiplication of them in this time. Thus 30,000 such mutations could essentially eliminate all the genetic diversity in the DNA if they were evenly spaced throughout the DNA and spread rapidly enough. They would have to spread in a total of about 400,000 years to produce the observed low genetic diversity. However, the fact that humans still have considerable genetic diversity (in blood types, for example) suggests that this mechanism has not been operating.
Also, having 30,000 such mutations, each with a fitness advantage of about .01, seems to be a large increase in fitness for a short (in evolutionary terms) time interval. In fact, .001 seems to be a more typical value for the selective advantage (according to Simpson, quoted in Spetner, Not by Chance, page 102), which would require 300,000 such beneficial mutations fixing in 400,000 years, a rather high rate of evolution. In reality, there would probably have to be many more than this to guarantee that the whole genome would be covered, due to their probable irregular distribution. If we assume one mutation fixing per year and 10 percent functional DNA, this leads to a rate of mutation of about one substitution per base pair per 300 million years. If we assume one percent functional DNA, this leads to a mutation rate of about one substitution per base pair per 30 million years. Both are much faster than assumed by evolutionists for functional genes, especially the latter rate. This makes such rates of evolution implausible, and strengthens the case that the human race (and probably all other species) are young.
Furthermore, many species persist for long (apparent) time periods in the fossil record with little change. These species do not appear to be evolving much, so we cannot use this mechanism of beneficial mutation spread to explain a low genetic diversity in them. Only a rapidly evolving species can reduce its genetic diversity in this way. If we try to escape from the dilemma by assuming that the rate of mutation is very small, explaining the small diversity, then the assumed evolution of species could not have occurred within the assumed time spans. The only plausible mechanism that can account for the data is that the human race experienced a severe population bottleneck recently which reduced the genetic diversity, or else was recently created essentially uniform. I suspect that the same reasoning applies to all other species, too, which seems to be a puzzle if one is not a creationist. In fact, this conclusion appears to flatly contradict the fossil record, which shows some species persisting with large populations for many millions of years. This is another problem for the standard theory, and suggests that the fossil record was laid down recently and quickly.
We also note that there is a tremendous difference between the kind of genetic diversity one would expect due to recent mutations, and that which is observed. If the existing genetic diversity among humans arose from recent mutations, then we should expect about 998 of 1000 humans to agree in a given base pair, and only about 2 in 1000 to differ. This should be true for all base pairs. The real situation is much different: for many alleles, there are several (or many, for some species) alleles that are quite common in the population. This is true for blood type, hair color, eye color, and so on. This seems to be proof positive that the human species did not reduce its genetic diversity by rapid evolution, nor did it generate it by recent mutations. All of these alleles would have had to originate recently and had a high selective advantage to become common in the population, yet none had a high enough selective advantage to eliminate the others. It must be the case, then, that the human race is young (or else there would be more diversity) and that the observed genetic diversity existed from the beginning. In fact, a very reasonable alternative hypothesis is that beneficial mutations are very rare and have only an insignificant effect on evolution, in most cases, even over large time periods. The observed changes are then almost all due to changes in frequencies of existing genetic material. Of course, this implies creationism.
Now, we discuss the fate of a mutation in the non-functional region of DNA. This part of the DNA is subject to recombinations (crossovers) which occur about once per 108 base pairs per meiosis. This means about 30 per generation for humans. These will tend to chop the genetic material up into pieces about 108 base pairs long. With each generation, the number of pieces per chromosome will increase linearly. If we have one crossover per chromosome per generation, then after one generation there will be two pieces, after two generations there will be three pieces, and so on. After a million generations the pieces will be 108 / 106 or 102 base pairs long on the average, each piece inherited by a different path. If man and ape separated 10 million years ago, then a million generations is a reasonable estimate. So we should not expect to see any piece of a pseudogene that has more than about 100 base pairs in it. Recall also that each pseudogene has only a small chance of fixing in the population, so that we should only see at most one piece of the pseudogene altogether.
The LGGLO pseudogene has 3364 base pairs, according to Edward Max. Thus if it arose all at once in some individual (unlikely), we would not expect to find more than about 1/30 of it in any human. This would not agree with the appearance of the whole pseudogene in one individual, which I believe was reported. It is more realistic to assume that this was initially a functional gene in all humans, which would be consistent with its common appearance. However, for insertions of new (neutral) material, this would not be so. The farther back one goes, the smaller the pieces should be. For a 20 million generations, the pieces of a pseudogene would be at most a few base pairs long on the average. This would correspond to possibly 50 million years, not a long period of time by evolutionary standards. We would never expect to find any common sequence having a hundred thousand base pairs, even for fairly recent events, unless such sequences were present in all individuals at some time in the remote past. This only pushes the problem back further in time, unless this material was beneficial to the organism at some time. Then it must have become beneficial to inactivate this genetic material for some reason, since neutral mutations will not spread, as a rule. Otherwise, it must be that life originated recently, or that there is a function for the "junk" DNA. It is also possible that a small population could lead to uniformity of the DNA.
If any evolutionist has answers to these points, I would most appreciate hearing from him or her. For example, one individual writes that single stretches of non-functional DNA will likely spread to the whole population eventually and in this way make all the non-functional DNA nearly identical. He feels that crossovers will not have much of an effect because the two chromosomes that are crossed over are already almost identical. I'd be interested to learn if others agree with this.
Here are some initial comments on this reply. It is true that according to population genetics, non-functional traits (and non-functional DNA) will eventually spread to the whole population. The expected time for this is 4N generations when it happens, where N is the population size. But in reality it should take much longer and may even be impossible, because one has sub-populations that tend to interbreed and are to some extent isolated from one another by geographic and social factors. And while this spread is taking place, the DNA is subject to being broken up by crossing over (recombinations). The respondent notes that it is estimated that all humans have a common paternal ancestor of about 40,000 years. This would be about 4,000 generations at most. This would correspond to a population size of about 1,000. But as I said, in reality this is probably not possible. Also, even if there were a population this small, due to the superior fitness of humans it would quickly increase, making such a spread of the Y chromosome unlikely. However, it is conceivable that a favorable mutation to the Y chromosome could rapidly spread to a larger population. One could then call the bearer of this mutation the first true human male. The common female ancestor is estimated at 200,000 years, which would correspond to 20,000 generations at most and a population size of at most about 5,000 but probably much less as argued above. It is difficult to see how a favorable mutation could lead to a more rapid spread of common female ancestry, but it is possible that a mutation to the mitochondrial DNA could accomplish this. Another plausible explanation is that at one time the only humans on earth were a man and a woman and their children, in the not too distant past.
Of course, we note that for evolution to proceed, one needs a large population size in order to have enough beneficial mutations to spread to the whole population. So one must have had populations in the millions (or billions) most of the time, implying a much slower rate of spread of neutral traits to the whole population and much more time for mutations and recombinations to enter in while this was happening.
We also note that some apparently non-functional traits such as being able to curl one's tongue and maybe blood type and eye color still show much variability within the human population, despite the tendency of genetic variability to reduce according to population genetics. So why should the non-functional DNA be any different? Why should we not see similar variability in the non-functional DNA? (The respondent points out that there is actually more variability in the non-functional DNA than in the functional DNA, though still not much.) And if these traits (such as eye color) are not neutral, the most beneficial ones should have won out even faster. In addition, it is a very strong statement to assert that all the DNA of all humans is very similar in the non-functional portion. By statistical arguments, we would expect some of the DNA to spread more slowly and some not to spread to all the population. To have all of the non-functional DNA in the whole population essentially identical is really extreme and a statistical impossibility, unless there was a dramatic population bottleneck as mentioned above. Since humans are probably not unique in this, one would have to assume similar bottlenecks for other species, too. Of course, this is not a problem for the creationist viewpoint. The other possibility is that this "non-functional" DNA really has a function.