SNP testing and genealogy

[TwiddlingThumbs]

I know you can use SNP testing to prove that two men DO NOT have a genealogical relationship, but can even Big Y and deep SNP testing show that people DO have a common male ancestor within the past 500 years? According to YFull, my current terminal SNP originated 1450 years ago. And, based on haplotype group studies, it seems that several other surnames share the SNP (which is not surprising since it originated before the start of surnames). It is therefore useless for proving that I DO have a relationship within the genealogical time frame with another male with my surname.

[MMaddi]

Originally Posted by TwiddlingThumbs

I know you can use SNP testing to prove that two men DO NOT have a genealogical relationship, but can even Big Y and deep SNP testing show that people DO have a common male ancestor within the past 500 years? According to YFull, my current terminal SNP originated 1450 years ago. And, based on haplotype group studies, it seems that several other surnames share the SNP (which is not surprising since it originated before the start of surnames). It is therefore useless for proving that I DO have a relationship within the genealogical time frame with another male with my surname.

The Big Y test, which tests about 10 million locations on the y chromosome, will find previously unknown SNPs. That's what it make it different from and superior to the testing technology in SNP packs. SNP packs are limited to testing about 160 already known SNPs specific to a particular branch.

The average mutation rate for new SNPs found in Big Y tests is estimated to be about 120 years. A similar test that covers more locations on the y, offered by Full Genomes Corp., has a mutation rate of about 90 years. So, Big Y can find new SNPs that occurred within the last 4 or so generations and the FGC test within the last 3 generations.

That's obviously well within the range of time since surnames came into common use among Europeans. The fact that YFull gives you a terminal SNP that's 1,450 years old probably means that they don't have many men in their database that share more recent common ancestors with you than that period of time or your haplogroup/subclade is uncommon and no other men in your subclade have taken the Big Y test.

Big Y, although expensive, can be used to break down a brick wall in the last few hundred years. The best strategy is to submit your unshared Big Y novel variants to YSEQ to have them made testable by development of primers for Sanger sequencing (another testing technology). Then if you have anyone with whom you think share a common paternal line ancestor around the time period of your brick wall, have him test your unshared novel variants at YSEQ at a very cheap price compared to a Big Y test.

[TwiddlingThumbs]

MMadi - Thanks for the explanation. Make sense. In comparing SNP testing and STR testing, would it be fair to say that, while both can be used to help establish a relationship within the genealogical time frame, the key differences are (i) SNP testing is much more definitive than STR testing, (ii) to get information relevant in the genealogical time frame, SNP testing is a lot more expensive than STR testing, and (iii) since there a lot fewer men who have taken the necessary level of SNP testing than have done STR testing, one is are currently less likely to find a match by SNP testing than by STR testing.

[MMaddi]

Originally Posted by TwiddlingThumbs

MMadi - Thanks for the explanation. Make sense. In comparing SNP testing and STR testing, would it be fair to say that, while both can be used to help establish a relationship within the genealogical time frame, the key differences are (i) SNP testing is much more definitive than STR testing, (ii) to get information relevant in the genealogical time frame, SNP testing is a lot more expensive than STR testing, and (iii) since there a lot fewer men who have taken the necessary level of SNP testing than have done STR testing, one is are currently less likely to find a match by SNP testing than by STR testing.

The answer to all your points is yes, especially your first point. Given that STRs can easily back mutate, they aren't as reliable as SNPs.

However, when you use the term SNP testing, the "yes" answer only applies to newly discovered SNPs found in Big Y tests. Results from SNP packs will not give you the resolution required to prove a common ancestor within the last few hundred years.

[MMaddi]

Here's a link - http://acreetree.net/ydnaadvanced.html - to a summary of an actual case study of how one surname project has used SNP testing to conduct their research and track branches from a common ancestor. These quotes from the press release are relevant to the questions you're asking:

"During the past two years, our project has been replacing conventional Short Tandem Repeat ('Y-STR') Y-DNA testing with Single Nucleotide Polymorphism ('Y-SNP') testing for most of our comparative efforts. This transition has greatly reduced the cost of testing after recovering an acceptable initial investment. Equally important, it has eliminated ambiguity in our comparisons....

"A persistent disadvantage of Y-STR testing has been that it requires laborious comparisons of lengthy strings of differing marker totals (called 'haplotypes') that can be indeterminate, even when a maximum number of markers are compared at substantial cost. Y-SNP testing, in dramatic contrast, permits unambiguous comparison of sequential, cumulative mutations. It is definitive, overruling any apparent Y-STR-based contradictions that may arise."

[spruithean]

Originally Posted by MMaddi

Here's a link - http://acreetree.net/ydnaadvanced.html - to a summary of an actual case study of how one surname project has used SNP testing to conduct their research and track branches from a common ancestor. These quotes from the press release are relevant to the questions you're asking:

"During the past two years, our project has been replacing conventional Short Tandem Repeat ('Y-STR') Y-DNA testing with Single Nucleotide Polymorphism ('Y-SNP') testing for most of our comparative efforts. This transition has greatly reduced the cost of testing after recovering an acceptable initial investment. Equally important, it has eliminated ambiguity in our comparisons....

"A persistent disadvantage of Y-STR testing has been that it requires laborious comparisons of lengthy strings of differing marker totals (called 'haplotypes') that can be indeterminate, even when a maximum number of markers are compared at substantial cost. Y-SNP testing, in dramatic contrast, permits unambiguous comparison of sequential, cumulative mutations. It is definitive, overruling any apparent Y-STR-based contradictions that may arise."

That is brilliant! But I would think Y-STRs would allow some clearer pictures after in-depth SNP testing to really determine the relationship between individuals.

[TwiddlingThumbs]

Originally Posted by MMaddi

It is definitive, overruling any apparent Y-STR-based contradictions that may arise."

Definitive, yes. But in this case, and in the case of most males with the same surname, there were no contradictions. The Acree STR results result in the same family groupings as are currently shown on their surname project web page.

[Mustik]

A quick question, and I apologise if this has been dealt with earlier, I have done SNP testing down through R-Z8 pack and ended up with an SNP below R-Z323 to R-CTS1747, from my understanding only one or two tested individuals have returned this SNP. Is this the end of the line? (I didn't have the funds to complete Big Y) Looking at the data, this mutation occurred several hundred years B.C. with no real geographical estimate as to location. Would Big Y have given me much more information?

[MMaddi]

Originally Posted by Mustik

Looking at the data, this mutation occurred several hundred years B.C. with no real geographical estimate as to location. Would Big Y have given me much more information?

A SNP pack is limited to testing for about 160 already known SNPs on a specific branch. In most cases, the most downstream SNP pack will get you down to a SNP that's about 2,000 years old, in some cases somewhat younger than that.

Big Y tests the y chromosome for about 10 million locations and is able to find previously unknown SNPs. Some of these SNPs will eventually match other men who take the Big Y test later or men who've taken a later version of a SNP pack to which previously unknown Big Y SNPs are added. That will get you down to SNPs that are 1,000 years old or sometimes less.

Also, Big Y will find SNPs that are unique to the last several generations of your paternal line, the last few or several hundred years. Those private SNPs may be useful in breaking down a genealogical brick wall. This is why haplogroup project administrators overwhelmingly recommend Big Y over SNP packs.

Big Y is on sale until Aug. 31 for $395. That's the lowest price it's ever had. Now's a good time to get the youngest SNPs possible through Big Y.

[TwiddlingThumbs]

Originally Posted by MMaddi

A SNP pack is limited to testing for about 160 already known SNPs on a specific branch. In most cases, the most downstream SNP pack will get you down to a SNP that's about 2,000 years old, in some cases somewhat younger than that.

Big Y tests the y chromosome for about 10 million locations and is able to find previously unknown SNPs. Some of these SNPs will eventually match other men who take the Big Y test later or men who've taken a later version of a SNP pack to which previously unknown Big Y SNPs are added. That will get you down to SNPs that are 1,000 years old or sometimes less.

Also, Big Y will find SNPs that are unique to the last several generations of your paternal line, the last few or several hundred years. Those private SNPs may be useful in breaking down a genealogical brick wall. This is why haplogroup project administrators overwhelmingly recommend Big Y over SNP packs.

Right, but if only 2 other people have tested at his current terminal SNP, chances are he will not match up with anyone at the lower SNPs. So, maybe useful in the future if and when a bunch of other people test, but until then?