Using deCODEme Results to Create a BGA Ancestry Test

[DKF]

I have been none too pleased that with all of those million SNPs, and at least 18,000 with ancestral informative value, that all those of us who have taken the deCODEme test have is raw data. Actually, not quite. There is a lot of information embedded in the output, but it is necessary to dig around to find it and make sense of it.

Thanks to the very clear suggestions of Anders Palsen, I found that it is possible to explore each chromosome in turn, looking for haploblocks that are attributable to European, African or East Asian (I suspect that this includes Native American to some degree). It is so simple - well only because Anders was able to master the procedural details. It boils down to learning how to use the browser that comes with out deCODEme results found by clicking on the "Advanced" section on the left side of the page.

I found an number of significant haploblocks of millions of base pairs in length that are apparently not from European sources. Since I have documentary evidence of Native American ancestry, and a strong suspicion of African ancestry (due to the phenotypes of earlier ancestors), but only on my Mom's side, this is much easier. Also since males have only one X chromosome, and it must have come from the mother, this is also of immense assistance.

In the next few posts I will outline the procedure and the findings. I would like to find out if folks who are 100% European also have some level of East Asian and African beyond background noise (i.e., on say chromosome 10, up to 50% East Asian and 50% European at points along the chromsome.

Ok, got to run now, but will be back to discuss this further with interested parties.

[DKF]

I have been none too pleased that with all of those million SNPs, and at least 18,000 with ancestral informative value, that all those of us who have taken the deCODEme test have is raw data. Actually, not quite. There is a lot of information embedded in the output, but it is necessary to dig around to find it and make sense of it.

Thanks to the very clear suggestions of Anders Palsen, I found that it is possible to explore each chromosome in turn, looking for haploblocks that are attributable to European, African or East Asian (I suspect that this includes Native American to some degree). It is so simple - well only because Anders was able to master the procedural details. It boils down to learning how to use the browser that comes with out deCODEme results found by clicking on the "Advanced" section on the left side of the page.

I found an number of significant haploblocks of millions of base pairs in length that are apparently not from European sources. Since I have documentary evidence of Native American ancestry, and a strong suspicion of African ancestry (due to the phenotypes of earlier ancestors), but only on my Mom's side, this is much easier. Also since males have only one X chromosome, and it must have come from the mother, this is also of immense assistance.

In the next few posts I will outline the procedure and the findings. I would like to find out if folks who are 100% European also have some level of East Asian and African beyond background noise (i.e., on say chromosome 10, up to 50% East Asian and 50% European at points along the chromsome.

Ok, got to run now, but will be back to discuss this further with interested parties.

Further to my earlier posting:

Procedure:

1) Sign on to decodeme.

2) Click on "Advanced" on the left hand side of the page.

3) Click on "Genome Browser again on the left.

4) Click the button that says "Launch Browser".

5) This may take a while (hopefully you have Java installed). Make sure you keep both windows open.

6) You will now have your browser window with all chromsomes at the top.

7) Start with chromosome 1 by clicking on it until it becomes darkened.

8) Under "Your ancestry" at the bottom (you may have to scroll down) you will see a straight or wavy line with yellow (East Asian) or Green (African). The rest (unseen) is European.

9) When you see an obvious point of interest, such as a triangle with a peak sticking well above the solid line put your mouse over the peak and see what percentages are involved (for each of the three biogeographical groups). Copy down the figure at the top of the peak.

10) Click above on zoom to spread out the peak and see where it begins and ends and copy down these figures (e.g., 35 million to 40 million base pairs).

11) The assumption is that this region may contain a haploblock from an ancestor who was not European. This is why I am curious as to what those from Europe would obtain when the do this.

Here are some of my results:

1) Chromosome 6:

A triangle beginning about 28 million bp to 33 million bp (so about 5 million bp haploblock).

European 68%, African 32%, East Asian 0%.

2) Chromosome 9:

A triangle from 26 million to 35 million (so a haploblocl of circa 9 million bp).

European 60%, African 0%, East Asian 40%.

3) X chromosome:

A triangle from about 50 to 65 million on the chart (about 15 million base pairs).

European .06%, African 99.04%, East Asian 0%.

Yes, that means 100% of this large region would likely be attributable to an African ancestor (at least that is the interpretation that appears to fit with the data).

Well, do others get similar or different results. Can we compare here so that all can better understand how we may use these test results as a tool to understand our autosomal intheritance.

David K. Faux.

[Maria_W]

I took DeCodeMe...

DKF.

Since I have Amerindian decendt I would also like to know what it might tell...I was really excited about the new DeCodeMe test. I could hardly wait to get my results back. But honestly now that I have them I am afraid I really don't know how to interpet them....I know that overall my Asian atDNA results were 6% and 9% for xDNA. So what does this mean?

Asian
Autosomal dna
Chromosome
4 is 11.7
9 is 10.8
13 is 10.0
14 is 9.6
12 is 8.9
11 is 8.2
5 is 8.1
8 is 7.9
6 is 7.7
1 is 6.7
10 is 6.3
20 is 4.7
17 is 2.5
19 is 1.9
3 is 1.8
7 is 1.5
15 is 1.4
16 is 1.2
2 is 1.0
21 is 0.9
18 is 0.5

Asian
x dna
x is 9%

Maria

[DKF]

DKF.

Since I have Amerindian decendt I would also like to know what it might tell...I was really excited about the new DeCodeMe test. I could hardly wait to get my results back. But honestly now that I have them I am afraid I really don't know how to interpet them....I know that overall my Asian atDNA results were 6% and 9% for xDNA. So what does this mean?

Maria, your NA ancestry, being from the 1600s, is going to be impossible to detect unless by some miraculous chance you inherited a haploblock that has been preserved for tens of generations. Just as a helpful way of looking at this, calculate the number of ancestors you have back to the 1600s - it will be in the millions. The decodeme test as it presently stands will be of no assistance - most unfortunately as it would be wonderful to have genetic confirmation of the documentary sources.

[DKF]

The decodeme test is a "bust" for BGA minority testing

The browser, via clicking on "Track List", allows you to compare your profile on each chromosome to your "friends", and as well to individuals from each of the reference groups.

I am not at all convinced that the output of this test "works". Those in my friends category born in Europe of European parents tend to have profiles not different in any way that I can see to those of us from North America. Each has locations on one or more chromosome where the percentage of East Asian and of African reaches say 40 percent or more. This statement applies equally to, for example, the French and Orcadian individuals that can be used for comparative purposes.

I then looked at the Mongolian and the Yakut individuals. Each was primarily East Asian on all chromosomes. Similarly with the Pima and Mayan of the Americas - mostly East Asian.

My conclusion is that there is far too much noise in this measure at present. The African and East Asian apparent haploblocks seen in European individuals is either a reflection of a tendency toward false positives; or this reflects very ancient ancestry.

deCODEme, in my opinion, needs to do some work on the ancestry part of their test and set the bar such that if one obtains African at 50% it actually means that this value confirms a haploblock inherited from an African ancestor. The fact that all East Asians and Native Americans were primarily "East Asian" suggests that the test cannot differentiate between the two groups.

Back to the drawing board.

David K. Faux.

[rucksack]

Bga

David, pardon me for butting in here, but what the heck is "BGA"?

I see no reference to it here, and I see no reference to it in scientific or genetic papers.

Thanks.

[DKF]

David, pardon me for butting in here, but what the heck is "BGA"?

I see no reference to it here, and I see no reference to it in scientific or genetic papers.

Thanks.

Biogeographical ancestry test, such as the DNA Print (ABDNA) test using 176 ancestrally informative SNPs to parse a person's genome into 4 Continental groups.

[tomcat]

One of the problems I see in your conclusion is that you are proposing multi-million bp LD haploblocks based on the single SNP's DeCode sequences.

If DeCode seqs a million SNP's across your 3+billion bp genome, your coverage on any single chromosome must be pretty scant. I don't know which X SNP's DeCode seqs but your presumably positive response within 15 million bp of your X (your example 3) cannot mean, indisputably, that you have a 15 million bp LD haploblock that is 99.04% African.

Might be better to focus in on the particular SNP to see the incidence of your alleles across the three different "ancestries."

Am not surprised that DeCode cannot differentiate East Asians from Native Americans. In a recent paper on European ancestry among Americans the authors were able to differentiate Ashkenazim from SE Europeans (with whom they usually consolidate) only through a Principle Component Analysis. Native Americans are, genetically, most like East Asians and, as a group, only unlike East Asians in their relative lack of diversity relative to East Asians. And as we know, the Native American haplogroups are better characterised as nested subclades to their Asian counterparts and Native Americans have the fewest private alleles relative to the major "ancestries."