...
EgyptSearch Forums Post A Reply
my profile | directory login | register | search | faq | forum home

» EgyptSearch Forums » Egyptology » Ancestry testing is bullshit » Post A Reply

Post A Reply
Login Name:
Password:
Message Icon: Icon 1     Icon 2     Icon 3     Icon 4     Icon 5     Icon 6     Icon 7    
Icon 8     Icon 9     Icon 10     Icon 11     Icon 12     Icon 13     Icon 14    
Message:

HTML is not enabled.
UBB Code™ is enabled.

 

Instant Graemlins Instant UBB Code™
Smile   Frown   Embarrassed   Big Grin   Wink   Razz  
Cool   Roll Eyes   Mad   Eek!   Confused    
Insert URL Hyperlink - UBB Code™   Insert Email Address - UBB Code™
Bold - UBB Code™   Italics - UBB Code™
Quote - UBB Code™   Code Tag - UBB Code™
List Start - UBB Code™   List Item - UBB Code™
List End - UBB Code™   Image - UBB Code™

What is UBB Code™?
Options


Disable Graemlins in this post.


 


T O P I C     R E V I E W
Ceasar
Member # 18274
 - posted
Hey guys,

I have posted this on forum biodiversity and I wanted to see what the ES crew take on it.

There seems to be numerous contradictions within the genetic framework. For example, a PCA plot. Supposedly they take hundred of thousands of SNP's and impose the data on the graph indicating the similarity between peoples obviously. You see the African Cline lead to the European Cline then the Asian Cline etc etc. Like peoples from certain populations are always going to be more similar to each other based on the PCA plot. But How about a paper like this

http://www.genetics.org/content/161/1/269

They found that Sub-Saharan Africans on average are more different from each other than they are to Eurasians. This is totality different then what you see with PCA and fst. How about this Pagani paper

https://www.ncbi.nlm.nih.gov/pubmed/26027499

It states that northeast african haplotypes were more similar to OOA populations (including Ethiopia's haplotypes) when they masked out non African ancestry (he considers non african to be the 40-50% arabic/levant admixture that shows in Ethiopians). This is implying that the sub-Saharan east African haplotypes are more similar to OOA then to west African. I have never seen this in any fst or PCA plot that indicates this. Even fst values fluctuate. I have seen fst values comparing similar populations between 4 to 10% (almost 3 times a difference). One last thing in this Tishkoff paper, she states that African and middle eastern share more alleles with each other than with any other population. But here African samples are not just restricted to north african.. the overwhelming majority of her samples are sub-Saharan African. But this does not show in PCA plot or fst. Middle easterners cluster with Europeans.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2947357/

You sometimes have Africans sharing more dna with Europeans then Europeans with Asians (which doesn't jive with fst or pca plot). I have seen this study somewhere but I can't find it. There is also a study out there which states that Craig Venter and James Watson (European scientist) share more SNP's with a Korean scientist then they do each other. Then you have people saying that when two people from different distinct populations will never been more similar to each other with enough markers.... which contradicts the Watson and Venter Study

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1893020/

I have other examples... but I don't need to post them. Also even with admixture charts when they break down in K's you can get some pretty different results based on the analysis and software. I remember when a user here posted that the abuser mummies had a large east African component that did not show up in the original paper. There is a henn paper where The Hadza and the sandwe are more different from each other (fst distance) then to Europeans (something that again is different on a pca plot) Even more recently I have seen a new paper on the Guanches (Berbers) were the authors got different results with different ancestry tools.

It seems like these scientist come up with different mathematical tools and assumptions about evolutionary history and then the shit bricks come out all contradicting each other. It seems like they are playing statistics with gene frequency's with limited data... how would these admixture calculators work if we had a reference of everyone dna? Would we even be able to accurately classify people?. Because it seems like the statistical power would shrink greatly because variation seems clinical. It just seems fucked up and I don't think this stuff is as solid as people make it out to be........
 
Elite Diasporan
Member # 22000
 - posted
Was it this thread?
http://www.forumbiodiversity.com/showthread.php?t=48968
 
Ceasar
Member # 18274
 - posted
quote:
Originally posted by Elite Diasporan:
Was it this thread?
http://www.forumbiodiversity.com/showthread.php?t=48968

Yeah, I am Taylor
 
Oshun
Member # 19740
 - posted
Still learning about DNA so I'll let ES have at the main response he's getting on FBD: You can still tell with accuracy if someone is European, Asian or African.

...assuming that statement is accurate, is that for populations who've been isolated or does consider admixture as well?
 
beyoku
Member # 14524
 - posted
Go ahead and spend the 100.00 to get tested. You will then seee how BS it is.
 
Oshun
Member # 19740
 - posted
https://www.youtube.com/watch?v=qyfWZZ7uPuE
 
capra
Member # 22737
 - posted
Different analyses measuring different things give different results.

quote:
Originally posted by Ceasar:
For example, a PCA plot. Supposedly they take hundred of thousands of SNP's and impose the data on the graph indicating the similarity between peoples obviously.

"Impose data on a graph" doesn't mean shit; what does PCA actually do?

quote:
They found that Sub-Saharan Africans on average are more different from each other than they are to Eurasians. This is totality different then what you see with PCA and fst.
What effect does high genetic diversity of populations have on their Fst and PCA placement?

quote:
It states that northeast african haplotypes were more similar to OOA populations (including Ethiopia's haplotypes) when they masked out non African ancestry (he considers non african to be the 40-50% arabic/levant admixture that shows in Ethiopians). This is implying that the sub-Saharan east African haplotypes are more similar to OOA then to west African.
If East Africans are closer to OoA than West Africans are, you say this means that East Africans are closer to OoA than they are to West Africans?

quote:
Even fst values fluctuate. I have seen fst values comparing similar populations between 4 to 10% (almost 3 times a difference).
What is Fst?

quote:
One last thing in this Tishkoff paper, she states that African and middle eastern share more alleles with each other than with any other population.
Where does she say that?

quote:
Also even with admixture charts when they break down in K's you can get some pretty different results based on the analysis and software.
What does ADMIXTURE do?
 
Ceasar
Member # 18274
 - posted
quote:
Originally posted by capra:
Different analyses measuring different things give different results.

Lol it seems that way

quote:
Originally posted by Ceasar:
For example, a PCA plot. Supposedly they take hundred of thousands of SNP's and impose the data on the graph indicating the similarity between peoples obviously.

"Impose data on a graph" doesn't mean shit; what does PCA actually do?

PCA shows the relationships between groups of populations?

quote:
They found that Sub-Saharan Africans on average are more different from each other than they are to Eurasians. This is totality different then what you see with PCA and fst.
What effect does high genetic diversity of populations have on their Fst and PCA placement?

I think that fst measures population difference in terms of allele frequency. Fst and pca would group sub-Saharan Africans close to each other. But that paper is saying that they are closer to eurasians than they are to each other. I'm I missing something?

quote:
It states that northeast african haplotypes were more similar to OOA populations (including Ethiopia's haplotypes) when they masked out non African ancestry (he considers non african to be the 40-50% arabic/levant admixture that shows in Ethiopians). This is implying that the sub-Saharan east African haplotypes are more similar to OOA then to west African.
If East Africans are closer to OoA than West Africans are, you say this means that East Africans are closer to OoA than they are to West Africans?

The Paganini paper states that the egyptian and Ethiopian haplotypes are closer OOA then they are to west Africans

https://www.ncbi.nlm.nih.gov/pubmed/26027499

quote:
Even fst values fluctuate. I have seen fst values comparing similar populations between 4 to 10% (almost 3 times a difference).
What is Fst?

fixation index?

quote:
One last thing in this Tishkoff paper, she states that African and middle eastern share more alleles with each other than with any other population.
Where does she say that?

http://science.sciencemag.org/content/324/5930/1035.full

She says it under "African variation in a world wide context"

quote:
Also even with admixture charts when they break down in K's you can get some pretty different results based on the analysis and software.
What does ADMIXTURE do?

There is a post that I posted on forum biodiversity where I listed additional evidence showing a pie chart where Europeans shared more snp's with west Africans than with Asians. A moderator point blank stated that could not be true and then proceeded to delete the post when I showed him the additional evidence.

Maybe I came off too strong with the title, but there seems like there are inconsistencies with the data, for example like IAM sharing more alleles with Yoruba than with naufitans but the PCA plot says other wise or maybe I am not reading it right

The deeper problem is this.. we have a tiny fraction of the DNA in the world and very little ancient dna. Assumptions about human migrations are being made and where stuff came from... especially when it comes to Africa. And worse it seems that human variation are on clines... how much harder will it be to distinguish between populations if variation is on a cline? Look at 23 and me... They can't even go to a 95% confidence interval. It stops at 90% and an enormous amounts of DNA becomes unassigned at that level.. at-least for me... think about how much harder it would be to assign someone properly if we have everyone's DNA in a database.
 
the lioness,
Member # 17353
 - posted
quote:
Originally posted by Ceasar:

There is a post that I posted on forum biodiversity where I listed additional evidence showing a pie chart where Europeans shared more snp's with west Africans than with Asians. A moderator point blank stated that could not be true and then proceeded to delete the post when I showed him the additional evidence.


So the evidence you presented was not derived from ancestry testing?
 
Ceasar
Member # 18274
 - posted
quote:
Originally posted by the lioness,:
quote:
Originally posted by Ceasar:

There is a post that I posted on forum biodiversity where I listed additional evidence showing a pie chart where Europeans shared more snp's with west Africans than with Asians. A moderator point blank stated that could not be true and then proceeded to delete the post when I showed him the additional evidence.


So the evidence you presented was not derived from ancestry testing?
No, it was allele sharing pie chart
 
the questioner
Member # 22195
 - posted
quote:
Originally posted by Ceasar:
They found that Sub-Saharan Africans on average are more different from each other than they are to Eurasians.

I remembered reading something like that

what are your opinions on Haplogroup E?
 
Ceasar
Member # 18274
 - posted
quote:
Originally posted by the questioner:
quote:
Originally posted by Ceasar:
They found that Sub-Saharan Africans on average are more different from each other than they are to Eurasians.

I remembered reading something like that

what are your opinions on Haplogroup E?

E defiantly comes from Africa

But i think that there have been some many migrations and mixing in the past.. a-lot more then we think..... look at what they found out about europeans... three different ethnic groups created them.... Most of their ancestry doesn't even come directly from Europe lol
 
Swenet
Member # 17303
 - posted
quote:
Originally posted by Ceasar:
There is a post that I posted on forum biodiversity where I listed additional evidence showing a pie chart where Europeans shared more snp's with west Africans than with Asians. A moderator point blank stated that could not be true and then proceeded to delete the post when I showed him the additional evidence.

What are the sources you're referring to?
 
Ceasar
Member # 18274
 - posted
quote:
Originally posted by Swenet:
quote:
Originally posted by Ceasar:
There is a post that I posted on forum biodiversity where I listed additional evidence showing a pie chart where Europeans shared more snp's with west Africans than with Asians. A moderator point blank stated that could not be true and then proceeded to delete the post when I showed him the additional evidence.

What are the sources you're referring to?
Here you go

https://images.nature.com/full/nature-assets/nature/journal/v467/n7319/images_article/nature09534-f1.2.jpg

Its from this paper

https://www.nature.com/articles/nature09534
 
Elmaestro
Member # 22566
 - posted
would you mind answering the questions Capra asked, except the one referring to the Tishkoff study?
I'd be more beneficial if we understood what you do and/or do not know and why exactly do you see these examples as contradictions or invalidating.
 
Ceasar
Member # 18274
 - posted
quote:
Originally posted by Elmaestro:
would you mind answering the questions Capra asked, except the one referring to the Tishkoff study?
I'd be more beneficial if we understood what you do and/or do not know and why exactly do you see these examples as contradictions or invalidating.

I did answer his questions.... its just in the black marking.....
 
Swenet
Member # 17303
 - posted
This is interesting data. You can see this in other data, too, if you know where to look.

I still have to read it again and think about how this translates to population affinity. But it's not a direct measure of population affinity, that's for sure. It doesn't mean that Europeans are closer to YRI than they are to the Asian sample. Also, this is not a final result as the SNP discovery process is ongoing. You can see this clearly in the fact that, among the novel SNPs, it evens out a bit more (i.e. there is much less of a discrepancy).

But thanks for the paper. I will use this in the future.
 
Swenet
Member # 17303
 - posted
quote:
Originally posted by Swenet:
This is interesting data. You can see this in other data, too, if you know where to look.

I still have to read it again and think about how this translates to population affinity. But it's not a direct measure of population affinity, that's for sure. It doesn't mean that Europeans are closer to YRI than they are to the Asian sample. Also, this is not a final result as the SNP discovery process is ongoing. You can see this clearly in the fact that, among the novel SNPs, it evens out a bit more (i.e. there is much less of a discrepancy).

But thanks for the paper. I will use this in the future.

Actually, I was looking at it wrong. It doesn't even out (the proportion stays the same). Europeans share >50% more SNPs with YRI than they do with East Asians, in both datasets.
 
the lioness,
Member # 17353
 - posted
quote:
Originally posted by Ceasar:
quote:
Originally posted by Elmaestro:
would you mind answering the questions Capra asked, except the one referring to the Tishkoff study?
I'd be more beneficial if we understood what you do and/or do not know and why exactly do you see these examples as contradictions or invalidating.

I did answer his questions.... its just in the black marking.....
No, the formatting is wrong everything is bold.
Please correct this in a new post or a
message to the moderator Elite Diasporan (click his name in a post of his) with the corrected formatting so he can replace it


Corrected Example:

quote:
Originally posted by capra:
Different analyses measuring different things give different results.


quote:
Originally posted by capra:

Lol it seems that way

For example, a PCA plot. Supposedly they take hundred of thousands of SNP's and impose the data on the graph indicating the similarity between peoples obviously.



 
the lioness,
Member # 17353
 - posted
quote:
Originally posted by NonFingo

All individual anthropological disciplines have these shortcomings (not just population genetics). This is why you do a multidisciplinary approach and independently verify your assumptions. For instance, if you have studied Guanches from different anthro disciplines, you know the denials in the recent paper of Guanches having additional Eurasian are bogus. If you do this long enough (and having investigative talent always helps) you will start to develop a feel for what is real in population genetics vs what is just an artefact or an author’s bias.




 
Elmaestro
Member # 22566
 - posted
@ceasar
Ok, it seems as if There's really one concept that you haven't grasped yet, which is Heterogeneity. what generally separates Africans from ALL other non Africans isn't the variation but rather the frequency.

working with smaller numbers, imagine 100 variants; 75 found among Africans 50 found in euros and another 50 found in east Asians. the Eurasians share 25 of them however they occur at an average frequency of 90% of euros and 80% of East Asians. Africans and Euros share 40 variants which occur in 90% of Euros but only 10% in Africans. Africans and East Asians share 30 variants but they're only frequent in 40% of Africans and 80% of east Asians.

Though it'll seem like Africans and Europeans should be the closest, Europeans and East Asians will be closer because at any given rate two random individuals have a much higher chance (72%) of sharing 1/4 of known variants that amount drops to about 0.14% of a random African and a Euro sharing 1/4 of these variants.
In very very very very very simplified terms those percentages represent the K's or shared membership in an admixture/structure run.

So you can see how Africans can share more Alleles with MeNa like tishkoff states, however MeNa and Europeans will be group closer due to allele frequency.

That paper you posted with Europeans sharing more alleles with West Africans is 7 years old, and was testing variants discovered at the time (there have been 22 official iterations since w/ more variants). But statistically the pattern remains solid. There were multiple instances of pulse migratory events indirectly from Africa to europe and vice versa. That pattern is even seen down to the most superficial level. But right now I believe most serious Afrocentered freelance researchers are looking at the great possibility that Africans themselves drifted at variable levels then recombined and/or migrated outside the continent eventually contributing to the West Eurasian genepool, particularly Europeans.
...which ultimately explains the frequency discrepancy.


...Btw No, Ancesty/genetics testing isn't bullshit... but it isn't "all-revealing".
 
the lioness,
Member # 17353
 - posted
 -
Tishkoff (2009)
 
Ceasar
Member # 18274
 - posted
quote:
Originally posted by Elmaestro:
@ceasar
Ok, it seems as if There's really one concept that you haven't grasped yet, which is Heterogeneity. what generally separates Africans from ALL other non Africans isn't the variation but rather the frequency.

working with smaller numbers, imagine 100 variants; 75 found among Africans 50 found in euros and another 50 found in east Asians. the Eurasians share 25 of them however they occur at an average frequency of 90% of euros and 80% of East Asians. Africans and Euros share 40 variants which occur in 90% of Euros but only 10% in Africans. Africans and East Asians share 30 variants but they're only frequent in 40% of Africans and 80% of east Asians.

Though it'll seem like Africans and Europeans should be the closest, Europeans and East Asians will be closer because at any given rate two random individuals have a much higher chance (72%) of sharing 1/4 of known variants that amount drops to about 0.14% of a random African and a Euro sharing 1/4 of these variants.
In very very very very very simplified terms those percentages represent the K's or shared membership in an admixture/structure run.

So you can see how Africans can share more Alleles with MeNa like tishkoff states, however MeNa and Europeans will be group closer due to allele frequency.

That paper you posted with Europeans sharing more alleles with West Africans is 7 years old, and was testing variants discovered at the time (there have been 22 official iterations since w/ more variants). But statistically the pattern remains solid. There were multiple instances of pulse migratory events indirectly from Africa to europe and vice versa. That pattern is even seen down to the most superficial level. But right now I believe most serious Afrocentered freelance researchers are looking at the great possibility that Africans themselves drifted at variable levels then recombined and/or migrated outside the continent eventually contributing to the West Eurasian genepool, particularly Europeans.
...which ultimately explains the frequency discrepancy.


...Btw No, Ancesty/genetics testing isn't bullshit... but it isn't "all-revealing".

I get that stuff varies by frequency... thanks for that but that that paper is still kind of an anomaly because they are testing for millions of alleles and you find that Europeans have more in common with africans than asians...

I have another example

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3366900/table/T2/

This person sequenced some mitochondrial DNA and found that Africans have more different from each other than they are to Eurasians... he even found that the Japanese and Chinese are more different to each other than some Europeans and a African group!

Then you have the scenario were Craig Venter and James Watson share less DNA with each other than they do with a Korean scientist.

A couple of statisticians found that around 10,000 years ago we would all have the same ancestors... this may or may not be true..

http://www.nbcnews.com/id/13621729/ns/health-health_care/t/we-all-have-same-ancestors-researchers-say/

I still think that this stuff as of current is extremely limited.... I feel that the admixture results would be completely different if we had access to everyone genome and every last bit of ancient DNA

What do you think?
 
Ceasar
Member # 18274
 - posted
quote:
Originally posted by Swenet:
quote:
Originally posted by Swenet:
This is interesting data. You can see this in other data, too, if you know where to look.

I still have to read it again and think about how this translates to population affinity. But it's not a direct measure of population affinity, that's for sure. It doesn't mean that Europeans are closer to YRI than they are to the Asian sample. Also, this is not a final result as the SNP discovery process is ongoing. You can see this clearly in the fact that, among the novel SNPs, it evens out a bit more (i.e. there is much less of a discrepancy).

But thanks for the paper. I will use this in the future.

Actually, I was looking at it wrong. It doesn't even out (the proportion stays the same). Europeans share >50% more SNPs with YRI than they do with East Asians, in both datasets.
I have another example

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3366900/table/T2/

This person sequenced some mitochondrial DNA and found that Africans have more different from each other than they are to Eurasians... he even found that the Japanese and Chinese are more different to each other than some Europeans and a African group!

There is also the situation with the paper that details Craig Venter and James Watson (white) being more different from each other than a Korean scientist

https://commons.wikimedia.org/wiki/File:Genetic_overlapp_venn_diagram_of_Venter,_Watson_and_Kim.svg

I just think that right now that we are making assumptions based on an extremely limited data set... I think that if we had everyone's DNA in the world and if we were able to extract DNA from every last human fossil the picture would look alot different......
 
Swenet
Member # 17303
 - posted
quote:
Originally posted by Ceasar:
quote:
Originally posted by Swenet:
quote:
Originally posted by Swenet:
This is interesting data. You can see this in other data, too, if you know where to look.

I still have to read it again and think about how this translates to population affinity. But it's not a direct measure of population affinity, that's for sure. It doesn't mean that Europeans are closer to YRI than they are to the Asian sample. Also, this is not a final result as the SNP discovery process is ongoing. You can see this clearly in the fact that, among the novel SNPs, it evens out a bit more (i.e. there is much less of a discrepancy).

But thanks for the paper. I will use this in the future.

Actually, I was looking at it wrong. It doesn't even out (the proportion stays the same). Europeans share >50% more SNPs with YRI than they do with East Asians, in both datasets.
I have another example

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3366900/table/T2/

This person sequenced some mitochondrial DNA and found that Africans have more different from each other than they are to Eurasians... he even found that the Japanese and Chinese are more different to each other than some Europeans and a African group!

There is also the situation with the paper that details Craig Venter and James Watson (white) being more different from each other than a Korean scientist

https://commons.wikimedia.org/wiki/File:Genetic_overlapp_venn_diagram_of_Venter,_Watson_and_Kim.svg

I just think that right now that we are making assumptions based on an extremely limited data set... I think that if we had everyone's DNA in the world and if we were able to extract DNA from every last human fossil the picture would look alot different......

This type of genetic information (mtDNAs) is unrelated to your topic. Your topic is about genome-wide ancestry and why different ways of measuring large amounts of markers give different results. As single chunks of inherited genetic material, haplogroups are completely different. They are just vestiges of ancient people we carry with us. They can tell you things like where you fit on the human tree in relation to ancient ancestors and the arrival of new people in a population. They cannot tell you how closely related you are to the next person.
 
Elmaestro
Member # 22566
 - posted
quote:
Originally posted by Ceasar:
I get that stuff varies by frequency... thanks for that but that that paper is still kind of an anomaly because they are testing for millions of alleles and you find that Europeans have more in common with africans than asians...

I still think that this stuff as of current is extremely limited.... I feel that the admixture results would be completely different if we had access to everyone genome and every last bit of ancient DNA

What do you think?

take your time and try to understand what I'm sayng about frequency vs shared alleles... things don't vary by frequency. We currently use frequency to determine closeness. the odds that random SINGLE individuals from populations A will share will share whatever amount of alleles also found in populations B determines closeness.

everything you seem concerned about, pans out.

You are 99% RIGHT when you say more people or different datasets can alter admixture results... but if you truly understood what I was explaining earlier, me telling you that ADMIXTURE simultaneously predicts the source (K) and the Frequency(membership) should make absolute complete sense.

You are aware that Africans are very diverse in comparison to Eurasians right? the amount of variants found in your own example even shows you that.

If I gave you 3 Africans and told you that each of them are closer to Iberians than they are to each other but they cluster with each other, understand that;
1- Whichever variants shared between the three Africans have an extremely high chance of being the same.
2- Each African share a random Different subset* of alleles with Iberians.

Work with even smaller numbers using the above example. out of 20 alleles the three Africans share 10, the same. exact. 10.... however each of them share 11 with a random Iberian. And Iberians on average share the same 15 alleles. How would they cluster w/ each other vs. Iberians?
 
capra
Member # 22737
 - posted
quote:
Originally posted by Ceasar:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3366900/table/T2/

This person sequenced some mitochondrial DNA and found that Africans have more different from each other than they are to Eurasians... he even found that the Japanese and Chinese are more different to each other than some Europeans and a African group!

READ the fucking sources you are citing, don't just pick out bits of undigested information. This is not even a study, it is an exercise for first-year students (a highly simplistic one too). The author did not sequence any DNA, he selected a small set of mtDNA sequences from an old study, giving 1 mtDNA sequence from each population. Any given population in reality will have many, many different mtDNA haplotypes. These are *not* comparisons between entire *groups*, but between 1 individual from each group.

Yes, some Africans are more different from some other Africans than they are from Eurasians - Eurasians are a subset of African genetic diversity. This is an incredibly basic piece of information, it is the genetic evidence for Out-of-Africa. How on earth can you possibly think this is a surprise?

The vast majority of non-Africans belong to haplogroup L3, specifically the M and N branches (some sub-branches of those are also found in Africa). There are 5 other L3 branches which are primarily African. The 3 Africans in this exercise who are closer to Eurasians belong to L3, which does indeed mean they are maternally more closely related to Eurasians than to the other Africans in the sample (who belong to L0, L1, and L2). The high distance between the Chinese and Japanese is the result of these particular individuals having a greater than average number of mutations (in fact they belong to different branches of M).

If you were to compare a whole set of mitochondrial haplogroups between Chinese and Japanese, and between Yoruba and English, you would find that the Chinese and Japanese are much closer to each other than the Yoruba are to the English, since Yoruba have about 2/3 haplogroups L2, L1, and L0, while Chinese and Japanese have all M and N with many more closely related branches.

Why don't you try learning the first thing about the subject before deciding it's bullshit?

for the human mtDNA tree:
http://www.phylotree.org

overview of African mtDNA:
http://www.isita-org.com/jass/Contents/2011Vol89/e-pub/21368343.pdf
 
the lioness,
Member # 17353
 - posted
Some people who say ancestry testing is bullshit prefer to go by phenotype because their polemic is built on that and most of the world is still playing by the old rules socially

The new paradigm, DNA, however also has potential to be misused
in new and more precise ways
But it is a deeper type of analyis, no doubt
 
capra
Member # 22737
 - posted
quote:
Originally posted by Ceasar:
PCA shows the relationships between groups of populations?

No, the kind of PCA we are talking about finds correlated alellic variation between *individuals*. It does not directly show relationships between groups, though you can to some degree deduce them from the results.

These principal components can (and usually do) reflect the contribution of ancestral groups with characteristic allele frequencies at different SNPs. But the choice of the PCs is dependent on who is in the analysis. The more individuals in the PCA share a particular set of genetic variants, and the more distinctive that set of variants is, the more importance the program will assign to that particular type of variation. Genetic drift, the random change in allele frequencies which causes distinct genetic signatures, is much stronger in smaller groups than in larger ones. Even though the splits are deeper between African groups, they have typically had large long-term population sizes relative to Eurasians, who first had the Out-of-Africa bottleneck, then very often additional periods of living as small isolated populations during the Ice Age.

So in a nutshell a PCA can be, and commonly is, dominated by younger but stronger genetic signatures over older but weaker ones. This is why a typical global PCA will first distinguish the Out-of-Africa genetic signature, then an East-West Eurasian one, and won't get to specifically intra-African differences until higher numbered PCs.

ADMIXTURE works differently - it assigns individuals to one or more hypothetical ancestral populations which would give the greatest likelihood of inheriting their observed genotype at each SNP - but it is dependent on the people selected for the analysis in much the same way as PCA is.

quote:
I think that fst measures population difference in terms of allele frequency.
Fst has several different definitions and methods of estimating it, so you have to be careful comparing Fsts from different studies as they may not be obtained in the same way. Broadly speaking Fst is a ratio of the variation *between* subpopulations to the variation "within* the combined population as a whole. Populations which have *different*, *small* subsets of the total genetic diversity have large Fst with each other. Populations which have *large* subsets of the overall diversity will have smaller Fst, because less of the variation is *between* the populations.

quote:
The Paganini paper states that the egyptian and Ethiopian haplotypes are closer OOA then they are to west Africans
I see no passage where it says that.

quote:
She says it under "African variation in a world wide context"
She says that East Africans and Middle Easterners share the most alleles *which are not shared with any other population*.

quote:
There is a post that I posted on forum biodiversity where I listed additional evidence showing a pie chart where Europeans shared more snp's with west Africans than with Asians.
Consider a SNP where CEU have 95% of the ref allele and 5% of the alt allele, CHB have 100% of the ref allele, and YRI have 95% of the alt allele and 5% of the ref allele. In that case CEU and YRI will share a SNP which CHB does not have, yet CEU is much closer to CHB than to YRI at that SNP.

Look, you don't know shit, OK? That's fine, everyone starts off not knowing shit. But don't go leaping in claiming it's all bullshit because you don't understand it. That's just annoying.
 
Ish Gebor
Member # 18264
 - posted
quote:
Originally posted by Oshun:
https://www.youtube.com/watch?v=qyfWZZ7uPuE

The Data Brokers: Selling your personal information

https://www.cbsnews.com/news/data-brokers-selling-personal-information-60-minutes/


23andMe Sells Data for Drug Search

https://www.technologyreview.com/s/601506/23andme-sells-data-for-drug-search/
 
the lioness,
Member # 17353
 - posted
quote:
Originally posted by Ish Gebor:
quote:
Originally posted by Oshun:
https://www.youtube.com/watch?v=qyfWZZ7uPuE

The Data Brokers: Selling your personal information

https://www.cbsnews.com/news/data-brokers-selling-personal-information-60-minutes/


23andMe Sells Data for Drug Search

https://www.technologyreview.com/s/601506/23andme-sells-data-for-drug-search/

Does this mean you would not or have not have your DNA tested?
 
Ish Gebor
Member # 18264
 - posted
^Hi, why didn’t you tell us what you would do/ whould have done.
 
the lioness,
Member # 17353
 - posted
quote:
Originally posted by Ish Gebor:
^Hi, why didn’t you tell us what you would do/ whould have done.

I was thinking about it but worry about private companies owning my genetic information
 



Contact Us | EgyptSearch!

(c) 2015 EgyptSearch.com

Powered by UBB.classic™ 6.7.3