PDA

View Full Version : The occurrence of DF27 subclades and where they are tested.



Earl Davis
08-21-2015, 08:28 AM
This thread is aimed at helping understand the currently known early subclades under DF27 including..

1. What the subclades are.
2. How common each subclade is
3. Which tests include which of these subclades

This work is flawed because

1. It's not finished
2. New early branches continued to be discovered
3. It looks at arbitrarily decided 'lead' snips only and the number of snips in each block vary
4. Some of these snips will move down the tree in the future as blocks of snips get split

For this exercise I call DF27 level 0 and then look at the next four levels under DF27. I chose 4 levels as you have to go that far down for some people to find their next down SNP that is not palindromic.

In looking at where snips are tested in panels I used the list of snips that was published when the test was first announced. In some cases snips may have been added to a panel at a later date.

This is certainly imperfect but it's aim is to highlight the subclades that exist and how common they are.

This used a sample of 290 kits on the Bigtree.

Several pics to follow this post in the next few mins.

Earl Davis
08-21-2015, 08:38 AM
This first picture shows levels, 0, 1 and 2

Including the palindromic snips and based on the sample of 290 kits 95% of people should find their level 2 SNP by taking a big Y. You can work out from this roughly what the percentage chance is of the various panels identifying peoples SNP at level 2. For example the DF27 panel at YSEQ might identify the level 2 snp for people in about 53% of cases.

5642

Earl Davis
08-21-2015, 08:46 AM
This next picture shows the third level under DF27. About 13% of people taking BigY do not currently find a match at this level but that's improving most weeks. I have not yet checked if all these snips are detected by BigY. At this point the majority of snips are not on current panel tests.

As well as the 13% of samples where the snp at this level is currently unknown a further problem is Z31644 and Z34609 accounting for another 19% of DF27 people between them. This group (which includes myself) need to go down to level 4 before that find a snp in a most desirable region of the Y.

5637

Earl Davis
08-21-2015, 09:01 AM
And so we reach level 4.

At this time about a quarter of the 290 kits can not be placed into a SNP at this level so this level still has much to reveal. This level is very much a moving feast at this time and several of the selected leads are in larger blocks that will soon split as new tests come in.

I am one of the 4 people out of 290 that find themselves in the FCG17112 group at this level. FGC17112 is one of a number of snips at this level where ALL the levels above it are problematic for one reason or another. In my case I go DF27>ZZ12>ZZ19>Z31644>FGC17112. The reason I mention this is that to include groups like mine some tests may need to look down as far as level 4 before they find a SNP they can used depending on the technology being employed. So for DF27 you can't just look at the first couple of levels and think you have most people covered off.

There could well end up being 50+ known snips at this level in years to come.

I have not completed the analysis of which tests cover the level 4 snips but I suspect most of them are not widely tested at this time outside of YElite and BigY.

5638

Earl Davis
08-21-2015, 09:10 AM
This tree diagram accompanies the above analysis and illustrates the branching points. It's based on the Bigtree but only includes levels 0-4 and only includes lead snips to keep it easy to follow. There may be a couple of cases where I have not been consistent in the lead snp between the tree and the spreadsheet screen shots but it is hopefully mainly consistent.

5639

It looks like you have to click on the image about 3 times to make it large enough to read easily.

Rick
08-21-2015, 11:07 AM
Thanks Earl. Very helpful. Please note the attachment on post #2 does not link properly.

Earl Davis
08-21-2015, 12:42 PM
Thanks Earl. Very helpful. Please note the attachment on post #2 does not link properly.

Thanks Rick. I have hopefully fixed the problem now. Please retry.

REWM
08-21-2015, 08:17 PM
Earl - FTDNA do not have a stand alone test for FGC32092 under Z2573. They have it listed as a Novel Variant in the Big Y. CTS1090 can only be tested at YSEQ and not at FTDNA.

Earl Davis
08-21-2015, 08:42 PM
Earl - FTDNA do not have a stand alone test for FGC32092 under Z2573. They have it listed as a Novel Variant in the Big Y. CTS1090 can only be tested at YSEQ and not at FTDNA.

Thanks. I had checked FGC32092 in FTDNA's SNP cart recently. I must have made a typo and put a 'Y' in that column rather than a 'N'. I did just double check right now and you are correct in noting that FTDNA does not currently offer it as a stand-alone test. I have corrected my local copy of the spreadsheet. In respect of CTS1090 I am not currently trying to properly keep on top of who is testing for what at 'level 4' until the tree is flushed out a little more at that level and stabilises a little more.

Earl.