Who here has done a GEDmatch test?

[land owner]

I wanna see what somalis would get on gedmatch

 

[Apollo]

I wanna see what somalis would get on gedmatch

GedMatch has a statistical noise problem. The results you get on there are not the same as those done on real time admixture analyses.

 

[land owner]

GedMatch has a statistical noise problem. The results you get on there are not the same as those done on real time admixture analyses.

Which company is the best when it comes down to breaking down somalis deep autosomal ancestry, I’m talking before cushitic formation deep

 

[Apollo]

Which company is the best when it comes down to breaking down somalis deep autosomal ancestry, I’m talking before cushitic formation deep

23andMe.com, if one scores between 98-100% Somali on there then one is simply pure Somali, especially if the remainder is Ethio-Eritrean or Sudanese (two noisy components close to the Somali one). Secondary analyses like GedMatch won't reveal much besides finding new relatives from other platforms like Ancestry.com who can also upload to it.

 

[land owner]

23andMe.com, if one scores between 98-100% Somali on there then one is simply pure Somali, especially if the remainder is Ethio-Eritrean or Sudanese (two noisy components close to the Somali one). Secondary analyses like GedMatch won't reveal much besides finding new relatives from other platforms like Ancestry.com who can also upload to it.

Is there any companies that’ll analyze the nilo and eurasian dna

 

[Apollo]

Is there any companies that’ll analyze the nilo and eurasian dna

It's included in the Somali component of 23andMe. Ethnic Somalis have the same levels of ancient admixture.

 

[MI]

Yeah, I’ve played around with the different calculators, mainly to see if it spotted an Indian component in my ancestry.

 

[MI]

Is there any companies that’ll analyze the nilo and eurasian dna

There’s several calculators that can roughly break down the different components in your ancestry. Here’s a guide for the different calculators:
http://genealogical-musings.blogspot.com/2017/04/finally-gedmatch-admixture-guide.html?m=1

 

[Apollo]

There’s several calculators that can roughly break down the different components in your ancestry. Here’s a guide for the different calculators:
http://genealogical-musings.blogspot.com/2017/04/finally-gedmatch-admixture-guide.html?m=1

GedMatch still gives you 2% noise levels and different results from actual admixture studies. I have tested it out myself. It is not reliable.

 

[MI]

GedMatch still gives you 2% noise levels and different results from actual admixture studies. I have tested it out myself. It is not reliable.

Yeah it’s not 100% but it doesn’t seem like suugo science either. It’s fairly consistent when it comes to how African/Eurasian shifted a person is, which is what OP seems interested in.

 

[Apollo]

Yeah it’s not 100% but it doesn’t seem like suugo science either. It’s fairly consistent when it comes to how African/Eurasian shifted a person is, which is what OP seems interested in.

No, I am not saying it is fake science, I am saying that it is erroneous.

GedMatch are based on previously run real-time admixture runs that did not include your allele frequencies. So when you include your own new sample in it the allele frequencies weren't known and you will get different results from your ethnic group who were included.

It's difficult to explain, but it is simply is not correct the figures you get on GedMatch, there is an up to 4% noise level. You can get components of which you actually have 0% ancestry from.

Beware the "calculator effect"

http://bga101.blogspot.com/2012/05/beware-calculator-effect.html

 

[MI]

No, I am not saying it is fake science, I am saying that it is erroneous.

GedMatch are based on previously run real-time admixture runs that did not include your allele frequencies. So when you include your own new sample in it the allele frequencies weren't known and you will get different results from your ethnic group who were included.

It's difficult to explain, but it is simply is not correct the figures you get on GedMatch, there is an up to 4% noise level. You can get components of which you actually have 0% ancestry from.

Beware the "calculator effect"

http://bga101.blogspot.com/2012/05/beware-calculator-effect.html

Apparently the Eurogenes and the MDLP K3b calculator have maneuvered around this problem don’t suffer from the same reference bias and run the reference sample the same way as the test sample? If I understand correctly