PDA

View Full Version : 23andMe archaic Y-DNA nomenclature conversion chart...does one exist?



Grossvater
03-27-2016, 04:47 AM
I'm very frustrated as an extremely amateur DNA researcher. My frustration comes from trying to compare 23andMe's archaic nomenclature with terms in current use for Y-DNA from ISOGG. Does anyone know of a chart online somewhere that converts the old to the new?

MfA
03-27-2016, 11:59 AM
Swedish Haplogroup Database: http://dna.scangen.se/index.php?show=tools&lang=en

ArmandoR1b
03-27-2016, 04:24 PM
The Swedish Haplogroup Database asks you to select your SNP but you have to get the SNP from the mutation mapper first https://www.23andme.com/you/labs/haplogroup_tree_mut_mapper/ or you can download the following table and look up the 23andme longhand name and get all of the other information - 23andme Haplogroups (https://www.dropbox.com/s/0lmocnq5b0epsh6/23andme%20Haplogroups%20SNPs%20and%20rsid.xlsx?dl= 0)

Heber
03-27-2016, 04:56 PM
23andme is still incredibly based on the ISOGG 2010 YTree.
You can see all the ISOGG YTrees here.
http://isogg.org/tree/index.html

You can then translate from the 2010 to 2015 YTree.

ArmandoR1b
03-27-2016, 07:17 PM
23andme is still incredibly based on the ISOGG 2010 YTree.
You can see all the ISOGG YTrees here.
http://isogg.org/tree/index.html

You can then translate from the 2010 to 2015 YTree.

Some of the 23andme haplogroups are from 2008. If you look at column E in the link I posted earlier with the table you can see that some match the 2008 ISOGG tree and some match the 2009 ISOGG tree. From 2009 to 2010 many didn't change which is why it looks like they are from 2010. The best way to convert is to look up the haplogroup in the mutation mapper first https://www.23andme.com/you/labs/hap...ee_mut_mapper/ then get the SNP then cross-reference the SNP with the 2016 ISOGG Tree, YFull, or the BigTree if the haplogroup is downstream from P312. This table (https://www.dropbox.com/s/0lmocnq5b0epsh6/23andme%20Haplogroups%20SNPs%20and%20rsid.xlsx?dl= 0)will also do the job though.

The 23andme haplogroups that don't match the 2009 and 2010 tree are G2a4, J2a1b2, J2a1b1a, J2a1a, J2a1a, J2a1b, J2a1b1, J2a1b1, J2a1b2, J2a1c, J2a1e, J2a1i, J2a1h1, J2a1j which are defined respectfully by the following SNPs. L32, M163, M327, M322, M47, M67, M260, M92, M166, M68, M319, P81, M137, P279

This is why it is important to identify the SNP using the mutation mapper or the table.

Grossvater
03-28-2016, 01:23 AM
Thank you, kindly to all who replied. The Swedish Haplogroup Database website gave me the answers I needed.

VinceT
03-28-2016, 03:39 AM
23andme is still incredibly based on the ISOGG 2010 YTree.
You can see all the ISOGG YTrees here.
http://isogg.org/tree/index.html

You can then translate from the 2010 to 2015 YTree.


Some of the 23andme haplogroups are from 2008. If you look at column E in the link I posted earlier with the table you can see that some match the 2008 ISOGG tree and some match the 2009 ISOGG tree. From 2009 to 2010 many didn't change which is why it looks like they are from 2010. The best way to convert is to look up the haplogroup in the mutation mapper first https://www.23andme.com/you/labs/hap...ee_mut_mapper/ then get the SNP then cross-reference the SNP with the 2016 ISOGG Tree, YFull, or the BigTree if the haplogroup is downstream from P312. This table (https://www.dropbox.com/s/0lmocnq5b0epsh6/23andme%20Haplogroups%20SNPs%20and%20rsid.xlsx?dl= 0)will also do the job though.

The 23andme haplogroups that don't match the 2009 and 2010 tree are G2a4, J2a1b2, J2a1b1a, J2a1a, J2a1a, J2a1b, J2a1b1, J2a1b1, J2a1b2, J2a1c, J2a1e, J2a1i, J2a1h1, J2a1j which are defined respectfully by the following SNPs. L32, M163, M327, M322, M47, M67, M260, M92, M166, M68, M319, P81, M137, P279

This is why it is important to identify the SNP using the mutation mapper or the table.




The Y-SNPs used by 23andMe were selected from the ones known in 2008. Its haplogroup assignments were later updated to ISOGG 2010, but its current chip design still only uses the 2008 era Y-SNPs.

ArmandoR1b
03-28-2016, 12:09 PM
The Y-SNPs used by 23andMe were selected from the ones known in 2008. Its haplogroup assignments were later updated to ISOGG 2010, but its current chip design still only uses the 2008 era Y-SNPs.

That's not adding up.

J1e exists at 23andme, the 2008 ISOGG tree, and the 2009 ISOGG tree but does not exist in the 2010 ISOGG tree. The defining SNP is P58 (rs34043621).

J2a1b1 exists at 23andme and in the 2008 ISOGG tree but does not exist in the 2009 or 2010 ISOGG tree. The defining SNP is M327 (i4000098).

The 2010 ISOGG longhand name for Q-M3 (rs3894) is Q1a3a1 but the 2008 longhand name is Q1a3a. The 2010 longhand name for Q-M19 (rs3910) is Q1a3a1a but Q1a3a1a does not exist at 23andme. The 2008 longhand name for Q-M19 is Q1a3a1 which does exist at 23andme.

The 2010 longhand name for G-L32 (rs7892988) is G2a3 but G2a3 does not exist at 23andme. The 2008 longhand name for G-L32 (rs7892988) is G2a4 and G2a4 does exist at 23andme.

These are all of the SNPs that don't match the 2010 ISOGG tree but do match the 2008 ISOGG tree. M86, M200, L32, M163, M327, P58, M322, M47, M67, M260, M92, M166, M68, M319, P81, M137, P279, M307.1, M3, M19, M194, M199, P106, M306, P98, M204, M56, M87, PK5, L4, P249, P267, M124, P77

VinceT
03-28-2016, 06:59 PM
That's not adding up.

J1e exists at 23andme, the 2008 ISOGG tree, and the 2009 ISOGG tree but does not exist in the 2010 ISOGG tree. The defining SNP is P58 (rs34043621).

J2a1b1 exists at 23andme and in the 2008 ISOGG tree but does not exist in the 2009 or 2010 ISOGG tree. The defining SNP is M327 (i4000098).

The 2010 ISOGG longhand name for Q-M3 (rs3894) is Q1a3a1 but the 2008 longhand name is Q1a3a. The 2010 longhand name for Q-M19 (rs3910) is Q1a3a1a but Q1a3a1a does not exist at 23andme. The 2008 longhand name for Q-M19 is Q1a3a1 which does exist at 23andme.

The 2010 longhand name for G-L32 (rs7892988) is G2a3 but G2a3 does not exist at 23andme. The 2008 longhand name for G-L32 (rs7892988) is G2a4 and G2a4 does exist at 23andme.

These are all of the SNPs that don't match the 2010 ISOGG tree but do match the 2008 ISOGG tree. M86, M200, L32, M163, M327, P58, M322, M47, M67, M260, M92, M166, M68, M319, P81, M137, P279, M307.1, M3, M19, M194, M199, P106, M306, P98, M204, M56, M87, PK5, L4, P249, P267, M124, P77

Sorry: to clarify, the Y-SNPs included on 23andMe's chip were not restricted to the ones registered by dbSNP or indexed by ISOGG. But they were identified as variants when the chip was designed circa 2008. SNP-chips aren't able to discover novel SNPs. They can only genotype previously known SNPs.

P58 was redefined from J1e in 2009 to J1c3 in 2010 by ISOGG. ISOGG noted in 2010 at the time that the "topology of P58, M365 and M390 markers has not been completely resolved." Consequently, 23andMe did not adopt this change.

23andMe has J2a1b1 mapped to M92 and M260, which was the case for ISOGG in 2008 and up to 16 April 2009, when Bonnie Schrack somewhat quizzically rearranged the ISOGG sub-tree under J2a-M410. Consequently, on 16 April 2009, M192 and M260 were reassigned from J2a1b1 to J2a4b1, and M327 was reassigned from J2a1b1a to J2a4b1a. Haplogroup J2a1 was left unassigned. These changes below J2a-M410 were not adopted by 23andMe.

In 2008, ISOGG had L32/S148 assigned to G2a4, but with the note that it was "confirmed within G2a, but exact location unknown", i.e. with respect to P16, M286, U8 and L30/S126, and L31. L32/S148 was revealed as synonymous with U8 in 2009, which defined G2a3. But oddly, 23andMe had left G2a3 undefined entirely, despite including L30/S126 (= rs34134567) in its raw data. U8 was not considered by the Karafet et al (2008) paper which defined the 2008 YCC tree.

M3 was redefined from Q1a3a in 2009 to Q1a3a1 in 2010 by ISOGG, when L53, L54, L55, L213 were added to the ISOGG tree above M3 on 29 July 2010. This occurred after 23andMe adopted the 2010 ISOGG tree. L53 (= rs34724285), L54 (= rs34954951), L55 (= rs35768544), L213 (= rs34549365) were added to the ISOGG tree thanks to Adrian Squecco's community analysis of 23andMe data. 23andMe has Q1a3a1 mapped to M19, which was the case at the end of 2009 and the first half of 2010.


Your insight indicates that 23andMe did not fully adopt ISOGG's 2010 Y-tree, or even ISOGG's 2008 and 2009 trees for that matter. But then it needs to be said that Y-haplogroups were never 23andMe's forte.

Little bit
03-28-2016, 08:37 PM
When queried directly to 23andme customer care, they officially state they are using ISOGG 2009 and PhyloTree Build 7. I copied and pasted this from an email answer from 23andme customer care:

8453

As the standard admonition says, 'results not typical, your results may vary.' I would not at all be surprised if 23andme is using a mishmash of ISOGG's and I remember them updating mtdna's in 2010 but build 7 was Nov 2009.

ArmandoR1b
03-28-2016, 10:53 PM
Sorry: to clarify, the Y-SNPs included on 23andMe's chip were not restricted to the ones registered by dbSNP or indexed by ISOGG. But they were identified as variants when the chip was designed circa 2008. SNP-chips aren't able to discover novel SNPs. They can only genotype previously known SNPs.

P58 was redefined from J1e in 2009 to J1c3 in 2010 by ISOGG. ISOGG noted in 2010 at the time that the "topology of P58, M365 and M390 markers has not been completely resolved." Consequently, 23andMe did not adopt this change.

23andMe has J2a1b1 mapped to M92 and M260, which was the case for ISOGG in 2008 and up to 16 April 2009, when Bonnie Schrack somewhat quizzically rearranged the ISOGG sub-tree under J2a-M410. Consequently, on 16 April 2009, M192 and M260 were reassigned from J2a1b1 to J2a4b1, and M327 was reassigned from J2a1b1a to J2a4b1a. Haplogroup J2a1 was left unassigned. These changes below J2a-M410 were not adopted by 23andMe.

In 2008, ISOGG had L32/S148 assigned to G2a4, but with the note that it was "confirmed within G2a, but exact location unknown", i.e. with respect to P16, M286, U8 and L30/S126, and L31. L32/S148 was revealed as synonymous with U8 in 2009, which defined G2a3. But oddly, 23andMe had left G2a3 undefined entirely, despite including L30/S126 (= rs34134567) in its raw data. U8 was not considered by the Karafet et al (2008) paper which defined the 2008 YCC tree.

M3 was redefined from Q1a3a in 2009 to Q1a3a1 in 2010 by ISOGG, when L53, L54, L55, L213 were added to the ISOGG tree above M3 on 29 July 2010. This occurred after 23andMe adopted the 2010 ISOGG tree. L53 (= rs34724285), L54 (= rs34954951), L55 (= rs35768544), L213 (= rs34549365) were added to the ISOGG tree thanks to Adrian Squecco's community analysis of 23andMe data. 23andMe has Q1a3a1 mapped to M19, which was the case at the end of 2009 and the first half of 2010.


Your insight indicates that 23andMe did not fully adopt ISOGG's 2010 Y-tree, or even ISOGG's 2008 and 2009 trees for that matter. But then it needs to be said that Y-haplogroups were never 23andMe's forte.

So in the end, each 23andme haplogroup needs to be looked up in the mutation mapper and cross-referenced in order to be sure that a person has the correct SNP since they don't always match up with the 2010 ISOGG tree.