View Full Version : Resolve Your Segments

08-16-2012, 11:24 PM
This is a quick intuitive write-up regarding a little process I have built on today, inspired by Davidski's The battle of the chromosome paintings: 23andMe vs. GEDmatch vs. SupportMix. (http://bga101.blogspot.com.au/2012/08/the-battle-of-chromosome-paintings.html)

All sorts of interesting coloured flecks can be found in my component-based chromosome paintings. However, certain regions have continued to hold much of my interest given their persistence across every calculator I've tried.

They definitely exist, since they're present across the board, but their ultimate origins are seldom answered based on the charts alone.

Are they indicators of recent mutual or derived ancestry with these regions? Or, do they represent prehistoric shared segments that cannot be expressed any other way by these programs?

The Process Outlined

The principle is simple;

1) Physically compare your chromosome paintings to date in ascending order of component specificity to isolate those persistent chromosome segments which stand out
2) Clarify the origin of these segments using your Eurogenes' SupportMix chromosome profile
3) Inspect the haplotypes for that chromosome to pinpoint precisely who you share the component with (if necessary?)

Note: To follow what has been described here, membership to the Eurogenes Ancestry Project (http://bga101.blogspot.no/) and GEDmatch (http://gedmatch.com/) are required.

Example 1: My East Asian Segments

All ancestry runs thus far have stated I have between 3.75-4.48% East Asian DNA, including components that are closely related (i.e. Siberian). This is one or two percentage points higher than the Iranian average, which may initially be explained by genetic inheritance from the Medieval Turks of Central Asia who arrived in West Asia circa 1000 AD.

I've decided to use my Chromosome Paintings from 23andMe and McDonald's BGA, as well as the Dodecad K12b and Eurogenes K13 runs via GEDmatch:

54 53 51 52

The East Asian segments confirmed by most of the paintings are shown below:

Chr1 - 2 segments (both East Asian)
Chr2 - 2 segments (Southeast Asian-East Asian, Siberian-Southeast Asian-East Asian)
Chr3 - 1 segment (pred. East Asian)
Chr8 - 1 segment (pred. Siberian)
Chr9 - 1 segment (unanimously East Asian)
Chr11 - 1 segment (unanimously Siberian)
Chr13 - 1 segment (pred. East Asian)

Below are the SupportMix results for those specific chromosomes:


On Chr1, one of the East Asian segments has switched to a Siberian-like Buryat_Tuvan component while the other has completely been replaced in its' general location by Chechen_Lezgin and Abkhazian.
Chr2 shows a similar trend with a Buryat_Tuvan and a Chuvash segment. It appears at this point that the positioning of segments here isn't congruent with what is shown on the chromosome paintings.
Chr3 is entirely West and South Eurasian. It isn't clear which segment matches the pred. East Asian one found on the paintings.
Chr8 matches the paintings exactly with a pred. Siberian segment right at the beginning. Curiously, there is a South Chinese component further along.
Chr9 also matches the paintints exactly, with the East Asian segment on those being represented by a South Chinese component again.
Chr10 continues the trend by also yielding a South Chinese segment.
Chr11 and Chr13 both also confirm the presence of East Asian derived segments, although they are the opposite of what was found on the chromosome paintings.

This exercise has confirmed that most of the East Asian derived segments in my autosomal DNA are indeed East Eurasian in origin. However, there was a fair bit of cross-talk between the segments that were predicted to be East Asian proper or Siberian on the chromosome paintings.

The picture of a near-evenly mixed East Asian-Siberian input into my genome is still consistent with it being Turkish in origin, given that is the case in Turkic-speaking groups further northeast (e.g Turkmen, Uzbeks, Kyrgyz, Altaians). (http://vaedhya.blogspot.co.uk/2012/03/north-european-component-variation.html)

I'm pleased with the findings here. Has anyone else had better success cross-validating any genetic peculiarities they've picked up during their personal genetics adventures?

08-30-2012, 04:54 PM

K=13 Results from Eurogenes

Chr painting, also K=13162