Admixture, as defined in genetics, is the identification of groups of SNPs that come from some earlier population. Likely isolated and interbred for several generations to make the "signature" unique to them. Admixture analysis is what results in the ethnicity pie charts popular with the genetic genealogy testing companies and mostly focuses on using the atxDNA results. This study is more concerned with recent regional origins than the haplogroup and phylogenetic trees that are more focused (traditionally) on ancient populations and migrations. But this focus is evolving. And the refinement of Admixture is becoming strong enough to get smaller, defined regions of interest to help with the study work.
This has been a developing science from population genetics simply because (a) there has not been widespread enough testing to truly uniquely identify clusters of SNPs to particular populations and (b) it relies on testing current individuals and surmising their genetic heritage / background. But great strides have been made of recent outside the academic setting by the genetic genealogy testing companies like Ancestry and 23andMe due to their large databases of testers. So much so that Ancestry is actually tracking clusters to immigrants and time periods in North American colonization. And 23andMe has been identifying DNA clusters down to not just a country but county districts within countries. Some of this was likely spurred by LivingDNA starting from some seminal research work on identifying clusters within U.K. counties.
We often have to remind people not to read too much into percentages below 5%. But sometimes not even that is enough given the variation we see. Again, capitalizing on our 7 siblings and their parents tested and highlighted on the Half Identical matching glossary entry, we show the Ancestral Regions or ad-mixture as reported by Ancestry. All 7 siblings and both parents were all tested on Ancestry. Due to a range of a few months between return of tests that coincided with a chip microarray change at Ancestry, some of the siblings results are from the Illumina GSA chip instead of the original Illumina Omniexpress one. Additionally, the siblings of the mother were tested (later. with the latest GSA version) and shown for comparison with the mother.
Notice how the mother shows only 94% Slovenian but her siblings show 100% and 98%. The parents are traced for many hundreds of years in the same location in Slovenia. The father of the 7 kids tested is 1/2 Lithuanian (Suwalki region which has some Polish influence) and 1/2 Slovak (near Hungary border and they identified as Hungarian). The children of that "94%" Slovenian mother have anywhere from 51% down to 46% for their reported Slovenian heritage. Remember, 50% of your DNA comes from each parent albeit the mother can appear to contribute more to sons if the X and Y are being counted for their size in base-pairs. The father shows close to the expected ethnic mix of 50% Lithuanian and 50% Slovakian. But the kids have wildly variating mixes from that paternal side. Anywhere from 34% down to 12% for Lithuania and 25% down to 8% for Slovakia. Oddly, we would expect the paternal grandmother (Slovakia) to have higher percentages because they usually overpower the paternal grandfather in providing more of their DNA to the grandkids. Often seeing a split of more like 30% from the grandmother and 20% from the grandfather. The drop from the father's expected regions is filled in mostly by various Polish regions and then a smattering of others. Oddly with two kids showing Sweden and one even some from Greece. Small amounts of Polish and Latvia were shown in the father but much larger percentages in the children.
This has been a developing science from population genetics simply because (a) there has not been widespread enough testing to truly uniquely identify clusters of SNPs to particular populations and (b) it relies on testing current individuals and surmising their genetic heritage / background. But great strides have been made of recent outside the academic setting by the genetic genealogy testing companies like Ancestry and 23andMe due to their large databases of testers. So much so that Ancestry is actually tracking clusters to immigrants and time periods in North American colonization. And 23andMe has been identifying DNA clusters down to not just a country but county districts within countries. Some of this was likely spurred by LivingDNA starting from some seminal research work on identifying clusters within U.K. counties.
Anecdotal study
|
Comparing siblings estimated Ancestral Regions |
We often have to remind people not to read too much into percentages below 5%. But sometimes not even that is enough given the variation we see. Again, capitalizing on our 7 siblings and their parents tested and highlighted on the Half Identical matching glossary entry, we show the Ancestral Regions or ad-mixture as reported by Ancestry. All 7 siblings and both parents were all tested on Ancestry. Due to a range of a few months between return of tests that coincided with a chip microarray change at Ancestry, some of the siblings results are from the Illumina GSA chip instead of the original Illumina Omniexpress one. Additionally, the siblings of the mother were tested (later. with the latest GSA version) and shown for comparison with the mother.
Notice how the mother shows only 94% Slovenian but her siblings show 100% and 98%. The parents are traced for many hundreds of years in the same location in Slovenia. The father of the 7 kids tested is 1/2 Lithuanian (Suwalki region which has some Polish influence) and 1/2 Slovak (near Hungary border and they identified as Hungarian). The children of that "94%" Slovenian mother have anywhere from 51% down to 46% for their reported Slovenian heritage. Remember, 50% of your DNA comes from each parent albeit the mother can appear to contribute more to sons if the X and Y are being counted for their size in base-pairs. The father shows close to the expected ethnic mix of 50% Lithuanian and 50% Slovakian. But the kids have wildly variating mixes from that paternal side. Anywhere from 34% down to 12% for Lithuania and 25% down to 8% for Slovakia. Oddly, we would expect the paternal grandmother (Slovakia) to have higher percentages because they usually overpower the paternal grandfather in providing more of their DNA to the grandkids. Often seeing a split of more like 30% from the grandmother and 20% from the grandfather. The drop from the father's expected regions is filled in mostly by various Polish regions and then a smattering of others. Oddly with two kids showing Sweden and one even some from Greece. Small amounts of Polish and Latvia were shown in the father but much larger percentages in the children.
External References
- Strider, Daniel, "Overview of Admixture Mapping", Curr Protoc Hum Genet. 2013 Jan; CHAPTER: Unit1.23. oi: 10.1002/0471142905.hg0123s76
- Khan, Razib, Analyzing ancestry with admixture, step by step, Discover Magazine, 14 Mar 2011 (utilizing the http://software.genetics.ucla.edu/admixture/UCLA Admixture tool)
- Wikipedia Genetic Admixture