Tribes scores

[kelly40]

Kelly,
My Ancestry by DNA test also showed a small amount of Sub-Sah. African (5%). Like yours, my Tribes test did not return results for any S-S Afric. country. Again, like you, I also received a fairly strong score for Mestizo in World Region Match. DNA Tribes defines "Mestizo" as Native American blended with European and African. So I think that my Mestizo result is reflecting my small percentage of African ancestry.
Did you suspect that you had any African ancestry before your Ancestry by DNA test? Mine was a complete surprise. My other results were: 85% Euro, 5% Native Amer., 5% East Asian, 5% Sub-Sah. African. Since some of my percentages are pretty small and possibly invalid, I'm planning to retake the Ancestry by DNA test when 3.0 (or whatever the next version is called) comes out.
Judy

After going through the tree info I do have, I had sort of expected some African, as both sides of the family came to this country in the 1600's, lived in areas where there was a lot of mixing of natives, africans and whites before it was deemed illegal and one branch that keeps leading back to NC to a mulatto family with same names, birth places and dates and one letter difference in the last name. I sort of got the feeling that they headed west, changing that one letter in the spelling and were able to pass as white. But of course have no way to prove it, at least not yet. I also read a slave narrative by a former slave (who evidently looked white) and used the same last name, of another side of the family saying that the family was okay with being indian, but was trying to arrange the marriages so that they would breed the "black" out as it had become too hard to be known as mixed. (not a direct quote, but the jist of it)
I do know that my father, now deceased, did not want me looking into the tree. He said that he didn't know why I wanted to "Dig all that **** up now". I am not sure what he was referring to. So, I guess the answer would be that nothing would really surprise me. I would be more surprised if my test came back 100% European.

Kelly

[tomcat]

You need to be computer literate to manage to run this yourself, but when first figuring how to do it, it seem simple to me. However so far I only have raw genotype data for tree populations all living in America and probably admixted to some degree. I need to solve the problem with decimal alleles and import of frequency data.

Noaide

Great that you got it to work or starting to work!

The LEA software also sounds interesting and useful. "The application of coalescent theory is revolutionizing standard population genetic analyses and admixture studies are no exception. Recently, a method based on coalescent simulations has been devised which enables drift in hybrid and ancestral populations to be incorporated into a likelihood-based estimation of admixture. This method, named LEA has the additional advantage of allowing the simultaneous estimation of the admixtures proportions and the time since the admixture event. The method has been recently applied to the issue of the relative Paleolithic and Neolithic contributions to European populations. (Chikhi et al "Y genetic data support the Neolithic demic diffusion model. Proc.Natl.Acad 2002 and also see Chikhi 2001 in Genetics for "Estimation of admixture proportions; a likelihood-based approach using Markov chain Monte Carlo").

from Jobling, Hurles and Tyler-Smith 2004

[Noaide]

Great that you got it to work or starting to work!

Here is a link to a collection of more genetic analysis software, some seem interesting to investigate more especially those involving invidual analysis vs populations.

I am still learning, I have added native american populations and all 4 main categories cluster quite nicely in each corner in the triangle plot and in the data plot when using 13 loci. I think it may be wise to weed out those individuals that show admixture to get a better model.

STRUCTURE has renewed my "hope" that CODIS can tell you something about your ancestry, however OMNIPOP is NOT the tool of choice to use!

Noaide

[burto]

You can research your alleles in combination. Each combination is a partial profile. Because every combination has a mirror or opposite combination this allows you to model and test partial profiles for your two parents.

When you say opposite or mirror combination, could you explain please?

[burto]

Here is a link to a collection of more genetic analysis software, some seem interesting to investigate more especially those involving invidual analysis vs populations.

I am still learning, I have added native american populations and all 4 main categories cluster quite nicely in each corner in the triangle plot and in the data plot when using 13 loci. I think it may be wise to weed out those individuals that show admixture to get a better model.

STRUCTURE has renewed my "hope" that CODIS can tell you something about your ancestry, however OMNIPOP is NOT the tool of choice to use!

Noaide

Which one did you go on? Could you explain how you did it?

[tomcat]

After receiving the results from the Tribes analysis shown below, I was surprised to get a result of 92% European and 8% Sub Saharan African result from Ancestry by DNA. Either could be correct considering the unknowns in the family tree, but it seems weird that they are so different. At least to me they seem to be. I don't have any Sub Saharan African in the Tribes report at all. Does anyone know if they have different definitions of what Sub Saharan Africa is?...

I got 9% Sub-Saharan African on AbDNA 2.5 that has never been confirmed by a better-quality Tribes population match, but is reflected in some family World scores: SSAF 0.04 and 0.3, West African 0.02, and Southern African 0.01. Some have pulled North African population matches and all have positive North African World scores.

(But then, despite our 50% Ashkenazi ancestry none of us have ever pulled a high-scoring match to any Jewish populations!)

So there is some support for the AbDNA reading in Tribes, and Sub-Saharan matches are found on Omnipop when working with individual alleles, marker-pairs and partial-profiles. AbDNA and Tribes could both be correct, the ancestry read by AbDNA admixture analysis doesn't read in Tribes profile-matching analysis, or to put it another way, due to our admixture our over-all profiles don't read Sub-Saharan to a significant degree.

We have gotten matches to admixed populations that have an African ancestral component but with everything else in the mix those matches don't provide much of a gauge.

[tomcat]

When you say opposite or mirror combination, could you explain please?

In this example with three markers you can only have one allele from each marker in any combination. So the other three alleles not in first combination form a combination that is the reverse of the first (they are opposites or mirror the first). If one combination is representative of one parent, it's opposite must represent the other parent.

[tomcat]

...
STRUCTURE has renewed my "hope" that CODIS can tell you something about your ancestry, however OMNIPOP is NOT the tool of choice to use!
...

Good! Omnipop does provide a handy db of pop stats if you can figure out a means of extraction other than re-keyboarding

[Noaide]

Good! Omnipop does provide a handy db of pop stats if you can figure out a means of extraction other than re-keyboarding

I wonder what makes the difference, I have made several runs trough structure 2.1 now both assuming independent alleles and correlated alleles but the result clearly put me with a probability over 90% that I do belong to the Caucasian group, the other tree group affiliation is close to 0.

Tom, as I remember you had some high NA scores in ABDNA, if will try to look for you codis alleles if you have posted them to see if there is any affiliation using this method.

Noaide

[tomcat]

I wonder what makes the difference, I have made several runs trough structure 2.1 now both assuming independent alleles and correlated alleles but the result clearly put me with a probability over 90% that I do belong to the Caucasian group, the other tree group affiliation is close to 0.

Tom, as I remember you had some high NA scores in ABDNA, if will try to look for you codis alleles if you have posted them to see if there is any affiliation using this method.

Noaide

You can find consolidated result for five siblings under thread 'Siblings and autosomal test (dna tribes)'. I am the first xy. The most likely tribal affiliation would be with Michigan NA, Minnesota NA and RCMP or CSFS Northern Ontario NA or NO Aboriginals.

HaplogroupC would be a good candidate too!

[tomcat]

Here is a link to a collection of more genetic analysis software, some seem interesting to investigate more especially those involving invidual analysis vs populations....

'Assignment' the most promising by description is for the recognition of origin populations of Canadian bears - Ursidae. Lots of fun!

There are some kinship apps available. The site is definitely worth a visit. There are online calculators (Web) and several platforms are supported (Mac, Win, Java, Unix etc).

[Noaide]

You can find consolidated result for five siblings under thread 'Siblings and autosomal test (dna tribes)'. I am the first xy.

Structure 2.1 thinks you are more European than me, 97.75% probability of belonging to Caucasian population. Much less than 1% probability you belong to any other group.

Enclosed barplot, yours the last one on the population RED to the right just before the African population barplot begins. I initially assigned you to the caucasian population, and the program confirm that this is the correct affiliation.

Please be aware that I just begun using this program yesterday, so things may change when I get more depth knowledge of the program.

Note the barplot diagram almost always group the individuals to the stated affilation, a few individuals have more admixture.

Noaide

[Noaide]

HaplogroupC would be a good candidate too!

Yes, but I cant find his alleles posted. As I can see he had 60% NA in ABDNA test so it should show something.

Noaide

[tomcat]

Structure 2.1 thinks you are more European than me, 97.75% probability of belonging to Caucasian population. Much less than 1% probability you belong to any other group.

Enclosed barplot, yours the last one on the population RED to the right just before the African population barplot begins. I initially assigned you to the caucasian population, and the program confirm that this is the correct affiliation.

Please be aware that I just begun using this program yesterday, so things may change when I get more depth knowledge of the program.

Note the barplot diagram almost always group the individuals to the stated affilation, a few individuals have more admixture.

Noaide

So each person is represented by a single bar and the height of bar indicates degree of affiliation - the shortest bars have greatest affiliation with background field? What do the bars mean that seem to descend from the top of the field - am I reading that right as descent?

I did get 60% IE on AbDNA.

No worries - I take all DNA results with a deep breath of air.

[tomcat]

Yes, but I cant find his alleles posted. As I can see he had 60% NA in ABDNA test so it should show something.

Noaide

You could PM her. She has her mother' alleles as well which might be even better!