Genetics and Human Brain Evolution

Eggasai, you are completely wrong here. The Nature article cites a total nucleotide divergence of around 4%; 1% caused by 35 million substitution mutations, and 3% caused by a total of 5 million indel mutations.You appear to have cherry-picked the lowest measure of identity you could find in the article (29%) in order to bolster your case. However that is the percentage of whole proteins (not nucleotide positions) which are identical in terms of their complete amino acid sequence.Just to repeat what the Nature article actually says: Around 96% of the total genome is identical (not 29% as you imply). These differences were caused by a total of 40 million mutation events (not 145 million, as you claim)Anyway, if the fact that 29% of whole proteins are identical in human and chimp seems low to you, you should spend a moment considering the implications. According to the Nature article, the remaining 71% of proteins differed by an average of only two amino acids. The average length of a protein is around 1000 amino acids. This means that the probability that an amino acid from a chimp protein is identical to the corresponding amino acid in the human orthologue is equal to : (0.29 * 1.0) + (0.71 * 998/1000) = 99.858%.If (as seems likely) the probability that human and chimp orthologous proteins are identical is inversely proportional to protein length, this estimate will be a little high. Let us imagine, conservatively, that the average of 2 amino acids differing per protein was consistent across all proteins, not just the longest 71% of them. This would mean that the probability that two orthologous amino acids are identical is equal to (998/1000) or 99.8%.So the average probability of identity for single amino acids is close to 100% and over 99.5%, whichever way you look at it. NOT 29%!!!Finally, there is your assertion that the natural mutation rate cannot account for the number of differences between chimp and human genomes. Let's say that the generation length for humans, chimps and their proto-species is around 15-20 years. Over the six million years since divergence, that gives us 300,000 to 400,000 generations per lineage. Since there are two lineages that can accumulate mutations, we have a total of 600,000 to 800,000 fertilization events separating a modern chimp from a modern human. Given that the nature paper declared 40 million mutation events, that gives us 50-66 mutations fixed per POPULATION per generation.The number of new mutations arising per individual is around 100. If we were to assume a mean historical effective population size of around 25,000 for each proto-species (consistent with chimpanzee demographic data), the observed divergence between chimps and humans requires something of the order of 2-3 mutations to be fixed for each 100,000 mutations occuring. That doesn't seem unreasonable by any means.Summary:

​

Probability that two orthologous amino acids are identical = 99.8%
Probability that two orthologous nucleotides are identical = 96%
Probability that two orthologous proteins are identical = 29%Edited by mick, : added estimate of per-amino-acid divergenceEdited by mick, : added thread title and summary at the endEdited by mick, : added mutation rate stuff

Hi eggasai, I can assure you that I am doing nothing of the sort. I read the pdf version of the paper yesterday evening, before replying to your post, and I have also read it in the past. I did not type anything into google, I located the article in an abstracting and indexing database and downloaded the pdf directly. The figures given in my reply were drawn directly from the article itself. Amusingly, immediately after telling me that I have confused the actual paper with the website announcing the paper, you go on to give a link to the website announcing the paper, and a quote from the website announcing the paper! WTF? I gave a quite a reasonably detailed albeit "back of the envelope" numerical analysis of the data and showed that the number of mutation events is perfectly consistent with natural mutation rates. If you have a problem with my numbers, you should tell me which ones you disagree with.In fact, I can provide direct quotations from the pdf to back me up.Here we go. I will give quotations from my previous post, and follow them with quotations from the pdf. The pdf is available here

quote:

---

We calculate the genome-wide nucleotide divergence
between human and chimpanzee to be 1.23confirming
recent results from more limited studies12,33,34. The differences
between one copy of the human genome and one copy of the
chimpanzee genome include both the sites of fixed divergence
between the species and some polymorphic sites within each species.
By correcting for the estimated coalescence times in the human and
chimpanzee populations (see Supplementary Information ”Genome
evolution’), we estimate that polymorphism accounts for 14-22% of
the observed divergence rate and thus that the fixed divergence is
1.06% or less... The analysis of modest-sized insertions reveals,32Mb of humanspecific sequence and ,35Mb of himpanzee-specific sequence, contained in 5million events in each species... On the basis of this analysis, we estimate that the human and chimpanzee genomes each contain 40-45Mb of species-specific euchromatic sequence, and the indel differences between the genomes thus total ,90Mb. This difference corresponds to 3% of both genomes and dwarfs the 1.23% difference resulting from nucleotide substitutions; this confirms and extends several recent studies63-67. Of course, the number of indel events is far fewer than the number of substitution events (,5 million compared with ,35 million, respectively)....
  —nature

---

So that's 5 million indels accounting for divergence in 3% of the genome, 35 million substitutions accouting for divergence in one percent of the genome, precisely as I stated. Clear enough? Anything in there that you specifically disagree with? Was my arithmetic incorrect? The assumption of generation time and population size reasonable?I think it is important to get the numbers right. I am happy to accept I have made mistakes if you care to show them, but assertions that I have not read the paper will not do.MickEdited by mick, : No reason given.

Hi eggasai,Your post was a little confused and I'm not really sure where to start. I don't really want to go through the whole thing so here are some of the more glaring errors: No. A single coding nucleotide sequence always has at least one reading frame. Whether fixed mutations are deleterious or not makes no difference to the proportion of observed differences between the two genomes. We don't even need to know the function of the genes that differ, just the number of nucleotides. I must say that your entire discussion of selection against deleterious mutation seems irrelevant when we are discussing fixed mutations. No. First, amino acids are not translated into proteins. More substantively, you have to read more carefully: They are saying that 29% of the proteins do NOT differ at any amino acid positions. The remaining proteins, which DO differ, differ by an average of only two amino acids. This is a surprisingly high level of similarity, not a low level.(in edit - I also need to read more carefully - the sentence is properly understood as saying that the median number of amino acid differences is two, including 29% of proteins with zero differences. An overall mean of 2 was used in my previous posts) Sure, I ignored chromosomal rearrangements since there are only nine of them. Compared to the 40 million substitution or indel mutations they are negligible in terms of the mutation rate calculations. Repeat my analysis with "forty-million and nine" mutations instead of "forty million mutations" and there should be little consequence. You think three nucleotides have to mutate in order to change a single amino acid? You are wrong. Hmmmn.... Hard to grasp what you are getting at, but you seem to think that only 20 possible codons are valid? This might explain why you think random mutation is such a non-starter. But if that's what you think, you are wrong. In the standard genetic code there are sixty-four possible codons (that is 4^3, not 4^4 as you claim) and arond 61 of them code for valid amino acids, the remainder for stop codons. all very nice, except that your "200 nucleotides" figure comes from a misunderstanding that the number of nucleotides that differ is equal to the number of necessary mutations that must be fixed. You can't seriously believe that as you have already given examples of single mutations resulting in numerous nucleotide differences (hint - a single deletion throwing off the reading frame).etc. etc.Edited by mick, : No reason given.Edited by mick, : No reason given.

Hi eggasai, It would be nice to go on to discuss the evolution of brain-related genes, as you suggest, but I don't see how it can be constructive if you misunderstand the most basic elements of molecular biology (notwithstanding your hubristic attacks on the understanding of others). How can you write that, directly after providing a figure showing precisely the opposite?Codons "code for" amino acids. A triplet of nucleotides (i.e. a codon) within the DNA sequence or within the transcribed RNA sequence directly corresponds to a single amino acid within the protein. That correspondence is called "coding". Amino acid sequences are not made out of codons, nor are codons "formed" in amino acid sequence. An amino acid sequence is just that - a series of amino acids. There are no nucleotides or codons within an amino acid sequence. No, the amino acid sequence IS the protein.I'm sorry to keep pressing on these basic matters but I don't see how it will be possible to discuss mutation rates in protein-coding genes involved in the brain without agreement on these matters.For example your belief: This is really just the result of your not understanding transcription and tranlsation. But your view that two out of three nonsynonymous mutations result in an inviable protein (presuming that a protein "not containing the amino acids of life" is inviable) is going to seriously cloud your judgement of the likelihood of mutations being fixed in brain-related genes.CheersMick Edited by mick, : No reason given.

Hi eggasai, You should bear in mind that brain size is correlated with body size. For this reason, absolute brain volumes are not a very useful comparative measure; rather we would normally want to use a relative measure of brain size which takes into account differences in overall body size. Because when the whole body gets larger, the brain gets larger as well but doesn't necessarily imply evolutionary change specific to the brain.If you take body size into account, it seems to me at least that to have the same size brain as an ape when you are only three feet tall is having a LARGE brain size, not a small one, since apes are aready considered fairly large-brained and they can grow up to six feet.The normal way to take account of body mass is to measure relative brain size as the observed brain size divided by the expected brain size. The expected brain size for any individual based on body size is given by the regression of log brain mass against log body mass. This measure is called the encephalization quotient (EQ). On this scale, a value of 1 means that you have a brain exactly as large as expected based on body mass. A value of two means that you have a brain twice as large as expected based on body mass. A value of 0.5 means that you have a brain half as large as expected, based on body mass. Etc. etc.If you calculate values for the various primates from fossils and extant species, and take the average EQ, you get the following values (mean from both sexes):

Species                           EQ
Homo sapiens (modern)             5.27
Homo sapiens (archaic)            3.87
Homo erectus                      3.27
Homo habilis                      2.73
Homo rudolfiensis                 3.13
Australopithecus boisei           2.14
Australopithecus aethiopicus      2.07
Australopithecus africanus        2.95
Chimpanzee                        1.39
Gorilla                           1.10
Gibbon (multiple species)         1.89-4.64
Guenons (multiple species)        1.45-2.49
Colobus monkey (multiple species) 1.08-1.21

SourceI think this data clearly shows that it is wrong to suggest, on the basis of absolute brain size in Homo habilis versus ape that no trend towards larger brains was present in the early members of the Homo genus.Apes are large animals, and they have large brains in the same way that they have large arms and legs, and large skulls. But gorillas and chimps still have larger brains than we would expect based on body mass (10-40% larger than predicted).The very earliest fossil hominids (Australopithecus) already had brains that were up to 100% larger than we would expect based on their body size, and the early members of the genus Homo had brains up to 227% larger than we would expect.If you look at these encephalizatoin quotients across time, you see a major jump in the Australopithecines, where go from an average kind of brain size to one that is twice as large as expected; then a steady, gradual increase within the Homo genus; followed by a second major jump in the early Homo sapiens, where we go from a species with a brain 3.9 times larger than we would expect based on body size, to a species with a brain 5.3 times larger than we would expect.CheersMickEdited by mick, : corrected a type - changed "200%" to "100%"Edited by mick, : No reason given.

A plot of absolute cranial capacity over time, data drawn from every hominin skull found prior to 2000, shows precisely the opposite. Whether you consider these skulls to be part of the genus Homo or part of another extinct group of apes, there is no decrease over time in absolute cranial volume. SourceNo, the skulls are placed in the "family tree" based on phylogenetic analysis of hundreds of morphological characteristics. These characteristics are strongly indicative of a greater affinity with Homo than with Pan or Gorilla. It's not a conspiracy or laziness - the shape of the tree is a mathematical result, it is not just drawn by hand from the imagination. You might make a valid argument that cladistic methods are inappropriate, but every proposed cladistic algorithm reaches a very similar conclusion, so you would have to propose a new algorithm.The species in the cladogram below that are labelled "Homo" are so simply because they cluster with Homo sapiens. But it doesn't matter whether you call them "Homo" or just "extinct apes", their biological relationship to humans is specified by the shape of the tree and not by the names we give to them. Source, based on 198 craniodental characteristicsThat's a good point, but since we don't have any bodies with tissues we have little choice but to infer body mass if we want to talk about these issues. A variety of hominid and ape body mass models can be used; they are able to predict (imprecisely but reasonably) the body mass of extant apes; and they all give pretty similar results in terms of the general trend in encephalization. One might also calculate an encephalization quotient based on the correlation between cranial capacity and some other measurable skeletal variable, but I can't find an example of that having been done and I foresee difficulties because craniodental characters will be functionally correlated as well as correlated by body mass allometry. You are still using absolute brain volumes, and this is statistically unsound because you are ignoring a known fact that brain size is a covariate of body size. Consider the body mass and endocranial volume of two Homo habilis individuals and the mean values for a few other apes.

Species              Body mass (g)    Endocranial volume (cm3)
Male Orangutan        87700           393.1
Female Orangutan      37800           341.2
Male Gorilla          169500          537.4
Female Gorilla        71500           441.4
Male Chimp            60000           406.5
Female Chimp          47400           370.6
Female Bonobo         33200           295.5
Female H habilis 1    30286           594
Female H habilis 2    34883           509

SourceAlthough the brain volume of female Homo habilis is of the same order of magnitude as the male Gorilla, the habilis body mass is about five times smaller. Similarly, although the body mass of female Homo habilis is of the same order of magnitude as the female Bonobo, the habilis brain volume is half as large again. What does this signify? Each of the species listed here has a limited supply of resources that it can invest in its own growth and development. Homo habilis is clearly investing an enormously larger proportion of resources than the male gorilla in brain growth as opposed to the growth of the body. It is this difference in the allocation of resources to the brain versus elsewhere that is of evolutionary interest. The absolute values are of little interest because natural selection is acting on the trade-off over allocation of limited resources, not on absolute magnitude. Sure, big animals have access to greater resources and can grow larger arms, legs, fingers and brains. But such an increase in absolute brain growth tells us little about the evolution and energetics of brain development. The fact that an African male human being has an absolutely larger brain than an Asian female, for example, does not tell us much of evolutionary interest, since African males simply tend to be larger than Asian females. Once you correct for size, the two groups are seen to invest rather similar levels of their energetic resources into brain development.If you think that correcting for body mass is irrelevant, why do you bring up the hominids rather than the dogs? They provide a far more striking example of rapid change in absolute brain size, with major differences such as between the brain size of a chihuahua and a great dane occurring over a few thousand years. Are such differences in dog brain size also impossible to account for with evolutionary theory? I doubt you'd even find a single creationist to agree with that.Cheers!MickEdited by mick, : typo

Let's try this one more time. Okay, this time I will ask you questions and I'll try to keep everything clear. In return, perhaps you could answer the questions directly and keep personal insults to a minimum. Question: what do you mean when you say "hundreds if not thousands of mutations in hundreds if not thousands of genes"? My understanding is that the number of mutations being tallied here was measured across the whole genome, not just in the genes. Even if the mutations were measured only in coding sequences (which I do not recall is the case) that would still be tens of thousands of genes, not hundreds. Are we not talking about 40 million and nine mutations in the entire genome?Question: under the assumption that mutations are distributed randomly, how many mutations per gene do you think these figures represent? Here are my calculations: About 1.5% of the human genome consists of protein-coding sequences (Reference). if the 40 million and nine mutations are distributed randomly, that gives us an expected number of 600,000 mutations in the coding portion of the genome. Since there are around 35,000 protein-coding genes in the human body (Reference), that gives us an expected number of 17 mutations per gene. Would you agree with this logic, and if not, why not? I know that the assumption of random distribution of mutations is not realistic, but I'd like to know if you agree with the ballpark figure of tens of mutations per gene rather than hundreds or thousands of mutations per gene. Question: why 20 mutations per year? You have said that there were forty million and nine mutations. Forty million and nine divided by seven million is 5.7, not 20.Question: what do you mean by "in their respective genomes"? The total number of mutations we are talking about here is the SUM of the mutations that have occurred either in the chimp lineage or in the human lineage. Just by comparing the chimp to the human, we cannot tell which lineage the mutations arose in. This is a minor point but I would like to know if you agree.After we've sorted out these basic facts, I have some more questions to ask, if that's okay with you.Thanks! I think short focussed posts like these will help the debate to move along nicely. I am happy to retract anything here if I've made a mistake so let's try to have a productive discussion.MickEdited by mick, : added pargraph starting with: Question: what do you mean when you say "hundreds if not thousands of genes"Edited by mick, : No reason given.

thanks, that's good to know. Lahn's paper was actually based on a comparison of human and macaque, which represents 20-25 million years of evolution. But eggasai is cramming this number of mutations accumulating in 20-25 million years into the 7 million years separating chimp from human.From the press release:

quote:

---

evolution of the human brain probably involves hundreds if not thousands of mutations in perhaps hundreds or thousands of genes -- and even that is a conservative estimate...It is nothing short of spectacular that so many mutations in so many genes were acquired during the mere 20-25 million years of time in the evolutionary lineage leading to humans

---

Well, that's one problem solved isn't it?MickEdited by mick, : added quote from news release

Hi eggasai,350,000 generations? It is 350,000 generations from the human to an unknown prehistoric ape, not 350,000 from the human to the chimp.Think about it. The common ancestor lived 7 million years ago. This means that 7 million years ago, two lineages diverged which since then have BOTH been accumulating mutations each generation. Let's assume that 20 years is a reasonable estimate for the historical generation time. The lineage leading up to chimps thus contains 350,000 generations-worth of mutations. And the lineage leading up to the human also contains 350,000 generations-worth of mutations. Hence when we compare the chimp to the human, we are talking about 700,000 generations-worth of accumulated mutations.This has been pointed out repeatedly over the course of this thread.MickEdited by mick, : No reason given.Edited by mick, : No reason given.