I need help understanding this quote about recent Natural Selection

[sevensages]

TL;DR Summary

Please help me make sense of the quote from Wade that Pritchard found 206 genetic regions under selection among Africans.

I am reading Nicholas Wade's book A Troublesome Inheritance. Please let's not make this thread a critique about the merits or demerits of the book. This thread is my attempt to understanding the evidence that Natural Selection in the human genome was recent and regional.

On Page 103 of A Troublesome Inheritance, Wade writes the following: "The regional nature of selection was first made evident in a genomewide scan undertaken by Jonathan Pritchard, a population geneticist at the University of Chicago, in 2006. He looked for genes under selection in the three major races--Africans, East Asians, and Europeans (or more exactly Caucasians, but European genetics are at present much better understood, so European populations are the usual subjects of study). Copious genetic data had been collected on each race as part of the HapMap, a project undertaken by the National Institutes of Health to explore the genetic roots of common disease. In each race Pritchard found about 200 genetic regions that showed a characteristic signature of having been under selection (206 in Africans, 185 in East Asians, and 188 in Europeans). But in each race, a largely different set of genes was under selection, with only quite minor overlaps.

The evidence of natural selection at work on a gene is that the percentage of the population that carries the favored allele of the gene has increased. But though alleles under selection become more common, they rarely displace all the other all alleles of the gene in question by attaining a frequency of 100%. Were this to happen often in a population, races could be distinguished on the basis of which alleles they carried, which is generally not the case. In practice, the intensity of selection often relaxes as an allele rises in frequency, because the needed trait is well on the way to being attained.

Geneticists have several tests for whether a gene has been a recent target of natural selection. Many such tests, including the one devised by Pritchard, rest on the fact that as the favored allele of a gene sweeps through a population, the amount of genetic diversity in and around the gene is reduced in the population as a whole. This is so because increasing numbers of people now carry the same sequence of DNA units at that site, those of the favored allele. So the result of such a sweep is that DNA differences between members of a population are reduced in the region of the genome affected by the sweep" .

______________________________________________________________________________________________________

I don't know if I fully understand the quote above.

Wade wrote that Pritchard found 206 genetic regions among Africans that showed a characteristic signature of having been under selection. Then Wade wrote that "The evidence of natural selection at work on a gene is that the percentage of the population that carries the favored allele of the gene has increased." Pritchard's study was published in 2006. I don't think that Pritchard scanned the human genome once in, say, the year 2000, and then Pritchard scanned the human genome a second time in 2006 and determined that the percentage of the African population that carried a favored allele of a gene increased from the year 2000 to the year 2006. How would Pritchard or anyone else know if the percentage of the population that carries the favored allele of a gene has increased?

Did Pritchard calculate that 206 genetic regions among Africans showed a characteristic signature of having been under selection because he found 206 genetic regions in which most Africans that had the favored allele of the gene, and most non-African humans did not have the favored allele of the gene in those exact same 206 genetic regions? In other words, did Pritchard discover 206 genetic regions among Africans that showed a characteristic signature of having been under selection by doing the following process:

Step 1# Pritchard found 206 genetic regions in which most Africans had the favored allele of a gene
Step 2# Then Pritchard discovered that most non-African humans (such as Europeans, Asians, American Indians, Australian aborigines) don't have the favored allele of the gene in those 206 genetic regions.

Is that how Pritchard determined that there are 206 genetic regions among Africans that showed a characteristic signature of having been under selection?

Please help me make sense of this.

 

[.Scott]

He is not claiming that there are only 206 such genetic regions - just that of the genetic regions he was looking at and of certain alleles of interest, only 206 showed apparent positive selection among the African populations in comparison to other populations he was looking at.

 

[sevensages]

He is not claiming that there are only 206 such genetic regions - just that of the genetic regions he was looking at and of certain alleles of interest, only 206 showed apparent positive selection among the African populations in comparison to other populations he was looking at.

I did not think that Pritchard is claiming that there are only 206 genetic regions under selection among Africans. Before you wrote post #2, I kind of figured that what you wrote in post #2 is the case.

Is my understanding of the two step process of how Pritchard found 206 genetic regions of selection among Africans correct?

How familiar are you with Pritchard's study that he published in 2006? Did Pritchard use tandem repeats or SNP's to find those 206 regions of selection among Africans?

 

[.Scott]

Rather than try to match up your words with the ones you quoted, I will give you what I get from that quoted section:
I'll start by coining the term "allele adoption" which I will define to mean the process of a specific allele "sweeping" through a population (as it is described in the article) to the point of being "well on its way to being attained" (as described in the article).

So the question being addressed is how can you tell if "allele adoption" has occurred recently or not so recently. And the method being considered is to examine the DNA region where that allele is found.
The section that you quote doesn't go much further than that. But the point would be this:
* Except for the adopted allele itself, if the DNA region around that allele shows a normal diversity, then that allele was not adopted recently. On the other hand, if that region show very little diversity, it is because that neighboring genetic information was just dragged around by the adopted allele recently.

I am completely unfamiliar with Pritchards study.
And since I am also unfamiliar with "A Troublesome Inheritance", you don't need to worry about me giving you a critique on that book.

 

[BillTre]

Not really aware of this book, but this is what it looks like is going on to me.
It seems to be a molecular study of the genome at a single given time. Not a two time point comparison.
By going through the genomes step by step, probably while comparing a "standard" sequence, places of sequence were identified.
Sweeps are generally identified by the neighboring sequences that are dragged along with the selected sequence. The more similar the adjacent sequences are, the more likely they were selected in a relatively short period of time (probably over several sequences).

This will happen for each of the identified regions around the selected gene. Since there are maybe 3-4 billion base pairs in the human genome, finding about 200 such sites in a genome is not surprising. All this work would be done on computers doing comparisons among the sequences of several different genomes. This involves lots of computer work. Biology has greatly benefited from increases in computer power and memory since genomes started getting fully sequenced.

The neighboring (or surrounding) sequences are selected because they are linked to the sequence being selected. Linkage means the selected gene is covalently linked on the same DNA molecule. Occasionally the linkage will be broken and the neighboring sequences could then become linked with a non-selected gene. Generally, breaking the linkage is more likely the more sequence is between two sites. It is much more likely to happen if the sequence is longer. (This is the basis of genetic mapping). As the genome goes through generations, other genes will then be linked to the selected gene. This will substitute other sequences to be dragged along through the generational selection sequence, reducing the likelihood of finding those sequences linked to the selected gene.
Sequences farther from the selected gene will be more likely to have their linkage broken and loss their ongoing association with the selected gene.
This process will get repeated over and over over several generations to give the results the study found.
The longer the time between a sweep and when it is sequenced, the less the neighboring sequences will look alike and the more recombination will substitute in more typical sequences.

 

[sevensages]

Rather than try to match up your words with the ones you quoted, I will give you what I get from that quoted section:
I'll start by coining the term "allele adoption" which I will define to mean the process of a specific allele "sweeping" through a population (as it is described in the article) to the point of being "well on its way to being attained" (as described in the article).

So the question being addressed is how can you tell if "allele adoption" has occurred recently or not so recently. And the method being considered is to examine the DNA region where that allele is found.
The section that you quote doesn't go much further than that. But the point would be this:
* Except for the adopted allele itself, if the DNA region around that allele shows a normal diversity, then that allele was not adopted recently. On the other hand, if that region show very little diversity, it is because that neighboring genetic information was just dragged around by the adopted allele recently.

What is the frame of reference for a normal diversity vs. very little diversity? For Africans, would you just compare the diversity in the DNA region around the allele in Africans to the diversity in the DNA region around the allele in non-African humans?

I am completely unfamiliar with Pritchards study.
And since I am also unfamiliar with "A Troublesome Inheritance", you don't need to worry about me giving you a critique on that book.

I am surprised someone as seemingly knowledgeable about genetics as yourself is unfamiliar with A Troublesome Inheritance. A Troublesome Inheritance is not an obscure book.

 

[BillTre]

A Troublesome Inheritance is not an obscure book.

Its obscure to me.

 

[sevensages]

Not really aware of this book, but this is what it looks like is going on to me.
It seems to be a molecular study of the genome at a single given time. Not a two time point comparison.
By going through the genomes step by step, probably while comparing a "standard" sequence, places of sequence were identified.
Sweeps are generally identified by the neighboring sequences that are dragged along with the selected sequence. The more similar the adjacent sequences are, the more likely they were selected in a relatively short period of time (probably over several sequences).

This will happen for each of the identified regions around the selected gene. Since there are maybe 3-4 billion base pairs in the human genome, finding about 200 such sites in a genome is not surprising. All this work would be done on computers doing comparisons among the sequences of several different genomes. This involves lots of computer work. Biology has greatly benefited from increases in computer power and memory since genomes started getting fully sequenced.

The neighboring (or surrounding) sequences are selected because they are linked to the sequence being selected. Linkage means the selected gene is covalently linked on the same DNA molecule. Occasionally the linkage will be broken and the neighboring sequences could then become linked with a non-selected gene. Generally, breaking the linkage is more likely the more sequence is between two sites. It is much more likely to happen if the sequence is longer. (This is the basis of genetic mapping). As the genome goes through generations, other genes will then be linked to the selected gene. This will substitute other sequences to be dragged along through the generational selection sequence, reducing the likelihood of finding those sequences linked to the selected gene.
Sequences farther from the selected gene will be more likely to have their linkage broken and loss their ongoing association with the selected gene.
This process will get repeated over and over over several generations to give the results the study found.
The longer the time between a sweep and when it is sequenced, the less the neighboring sequences will look alike and the more recombination will substitute in more typical sequences.

Your post seems to confirm what Scott wrote in post #4 on this thread. Do you agree with everything that Scott wrote on post #4 on this thread?

 

[sevensages]

@.Scott
@BillTre

Based on your posts, it appears to me that I am right about the two step process I described in the OP.

 

[BillTre]

Your post seems to confirm what Scott wrote in post #4 on this thread. Do you agree with everything that Scott wrote on post #4 on this thread?

yes.

 

[BillTre]

@.Scott
@BillTre

Based on your posts, it appears to me that I am right about the two step process I described in the OP.

I am not convinced that they are different steps. They could easily be done simultaneously by a computer.

 

[sevensages]

Please look at this photograph of a diagram of the results of Pritchard's research. Notice Pritchard found ten regions of the genome that are highly selected in ALL three of the major races. I don't understand this. If a region of the genome is highly selected for in ALL three of the major races, who is the frame of reference? American Indians and Australian Aborigines?

If ALL three of the major races (caucasian, Asian, and African) have the same 10 alleles favored, how would it be accurate to say that those ten alleles are favored? I would estimate over 98% of the entire human population is Caucasian, Asian, or of African descent.

I thought that the definition of a genetic region favored by one races is that most of the human population has a different allele at that genetic region.

 

[Hornbein]

It appears the idea is that if the diversity of the allele and its surroundings is low then it was recently selected for. That makes sense. It could be that it has been selected for in all three races.

 

[sevensages]

It appears the idea is that if the diversity of the allele and its surroundings is low then it was recently selected for. That makes sense. It could be that it has been selected for in all three races.

That is a mighty pithy post. You might have completely explained away the mystery to me of the diagram on page 104. Thank you.

 

[sevensages]

It appears the idea is that if the diversity of the allele and its surroundings is low then it was recently selected for. That makes sense. It could be that it has been selected for in all three races.

Wait a minute.

Wouldn't it be more accurate to write the following in the first sentence:

"It appears the idea is that if the diversity of the surroundings around the allele is low, then it was recently selected for."

Those ten alleles are the same alleles in all three of the major races. So I don't see why you would include the diversity of the allele itself in that sentence.

 

[Hornbein]

Wait a minute.

Wouldn't it be more accurate to write the following in the first sentence:

"It appears the idea is that if the diversity of the surroundings around the allele is low, then it was recently selected for."

Those ten alleles are the same alleles in all three of the major races. So I don't see why you would include the diversity of the allele itself in that sentence.

Maybe I'm missing something but I don't see how it makes any difference if it is the allele of main interest or the neighboring ones. The "age" of the allele is measured by how many mutations appear. The more mutations, the longer the allele has been around to accumulate them.

 

[sevensages]

Maybe I'm missing something but I don't see how it makes any difference if it is the allele of main interest or the neighboring ones. The "age" of the allele is measured by how many mutations appear. The more mutations, the longer the allele has been around to accumulate them.

You're not missing something. I made a mistake in post #15.

 

[sevensages]

It's my understanding that there are two ways to determine if an allele has been under recent selection in a racial group:

1# Comparing the frequency of the allele in a certain racial group to the frequency of the allele in the rest of humanity outside that racial group. i.e. If an allele is carried by 70% of people of European descent, and that same allele is only carried by 30% of humans that are not of European descent, then the allele has been under recent selection in Europeans.

2# Checking for a lack of diversity of the SNP's of an allele and its surroundings. Throughout most of their lengths, the genomes of humans are diverse in terms of the SNP's that they carry. But in some places the diversity suddently fades away, and there are large numbers of people carrying the same long identical blocks of DNA. Somewhere in the block is the advantageous gene which has dragged the whole block of DNA with it. Reduced diversity is a signature of recent selection. It has to be recent because over the generations the blocks get whittled down and fade from view.

_________________________________________________________________________________________________

Number 1# (comparing the frequency of an allele in one racial group to the frequency of the allele in the rest of humanity) is only used to determine if an allele is under selection in a certain individual racial group or more than one racial group, but not all of humanity. Number 1# cannot be used to determine if an allele was selected for in all of humanity because all 1# is doing is comparing frequency of alleles from one racial group to another, not comparing the frequency of an allele at two different points in time.

2# (checking for a lack of diversity of the SNP's of an allele and its surroundings) is the method that Pritchard used in his 2006 study. 2# can find if an allele has been under recent selection for all of humanity. But 2# can also be used to see if an allele has been under selection in an individual racial group only.

@Hornbein Do you agree with everything I wrote in this post? If not, what do you disagree with?

 

[BillTre]

I don't think that #1 works.
The presence of an allele in one population and not another is not evidence of it being under selection.
It could be due to random drift. There are probably other explanations.
Not every allele has a plus or minus selective value and not all selective values are strong enough to have an effect. Population size also has influences this.

 

[sevensages]

I don't think that #1 works.
The presence of an allele in one population and not another is not evidence of it being under selection.
It could be due to random drift. There are probably other explanations.
Not every allele has a plus or minus selective value and not all selective values are strong enough to have an effect. Population size also has influences this.

First of all, 1# is about comparing the frequency of an allele in one racial group to the frequency of an allele in the rest of humanity outside of that racial group, not about the existence of an allele in one population and not another. The details matter.

The racial groups that I am writing about comprise hundreds of millions of people. I don't think that random drift would be a decisive influence in large groups of hundreds of millions of people. Natural selection would be decisive in such large groups. Random drift could only be decisive in small societies.

You seem to disagree with Nicholas Wade. Wade wrote the following: "The evidence of natural selection at work on a gene is that the percentage of the population that carries the favored allele of the gene has increased" (103).

 

[BillTre]

First of all, 1# is about comparing the frequency of an allele in one racial group to the frequency of an allele in the rest of humanity outside of that racial group, not about the existence of an allele in one population and not another. The details matter.

Its better to talk about populations rather than racial groups. Its less racial.
If a population is founded with a particular allele frequency which is not under selection, thanthe frequency will not change much.

The racial groups that I am writing about comprise hundreds of millions of people. I don't think that random drift would be a decisive influence in large groups of hundreds of millions of people. Natural selection would be decisive in such large groups. Random drift could only be decisive in small societies.

If an allele involves the difference in one nucleotide in a codon for an amino acid coding for a protein, but the change in nucleotide does not result in a change of the proteins amino acid sequence, most people consider this to be selection free. It is possible that using different tRNAs would have some effect, but it would likely be below the threshold for being selected.

You seem to disagree with Nicholas Wade. Wade wrote the following: "The evidence of natural selection at work on a gene is that the percentage of the population that carries the favored allele of the gene has increased" (103).

I don't disagree with that statement, but I don't think that there is any direct evidence of changes in this case.
Just a current static condition about which you want to derive some conclusion. The increase of an allele has not been shown, but inferred.