The Clades Defined

The basic W - Wilma, the nominal ancestor of all Haplogroup W's - is descended from Haplogroup N and defined by the 1243C change in the coding region (the 8994A change has also been used to assign individuals to Haplogroup W). In HVR1 the basic W motif was 16223T, 16292T and 16519C, and in HVR2 073G, 189G, 195C, 204C, 207A, 263G, 309.1C, and 315.1C. There are two alternate clade definition systems for W's, depending whether one only takes into account coding region changes, or also include certain HVR changes that seem to be very stable. This is because there seems to be a great divide in the individuals tested to date between those with a 194T change in HVR1 and those without. There is also a group of individuals that have 'lost' the 16292T change in HVR1. Taking only changes in the coding region, the clades would look like this:

• W1 - defined by a 7864T change in the coding region. In HVR2 the lack of the 194T change in HVR2 is 99% correlated with W1 as shown by the comparison of unique HVR1 motifs indicated as non-W3 in the Genographic results with the same motifs in individuals that had HVR2 tested. The exceptions are noted below.
• W2 - was assigned in error in an early paper and is not used.
• W3 - defined by changes 1406, 13263, and 15784 in the coding region and accompanied by 194T in HVR1. 13263 is tested by FTDNA and Genographic since it is diagnostic (without the 1243 or 8994 changes) of Haplogroup C. W individuals with the change at 13263 can be assigned by definition to W3.
• W4 - defined by changes at 6528, 10097, and 15775 in the coding region and accompanied by 194T in HVR1 and 16362C in HVR1. This clade was identified by Peter Fischer. Of the two full-sequence W4's so far, both had the 16362 change in HVR1. Furthermore, Genographic results shows that no individual with 16362 tested as W3; and HVR2 results confirm that no individual with 16362 had the loss of the 194T change diagnostic of W1. Therefore, any individual with the 16362 change may tentatively be assigned to the W4 clade. This is subject to change as more full-sequence information becomes available.
• W5 - preliminary definition. There is one full sequence, with a change at 8860 as well as a 16234 change in HVR1. This 16234 change is also present in a number of individuals who tested as non-W3 by Genographic, and also with the presence of 194T in HVR2 (i.e., not W1). They are also assigned to this clade.
• W6, W7, W8 - these are tentative designations for three unique individuals characterized by coding region changes at 12215, 5147, and 5580, respectively. Two had the 194T change but the last one did not have HVR1 tested.

143A - 192C - 194T - 196C

There is an absolutely distinctive motif in HVR2 (143A - 192C - 194T - 196C). Bizarrely, this motif occurs at five different places in the W descent tree as clearly otherwise defined by coding region changes: at W1a, W1c, W6 and W7 (W7 lacking the 192C). This remarkable sequence cannot be due to shared ancestry, since there are multiple distinctive coding region changes separating the individuals. They cannot be due to some common laboratory problem or contamination, since in 8000+ mitosearch individuals they occur only in W individuals, and the same motif has been detected in the HVR2 by a number of different test services. Is there something in the DNA copying machinery of W's that repetively results in this same cluster mutation?

Loss of 16292

16292T is the marker for Haplogroup W which differentiates it from Haplogroup N. However in about 12% of W's it is lost. This is not a kind of post-N, pre-W, since coding region mutations show these occur well down the descent tree. This has occurred at least three times: in W1c, W3, and a strange group of East Asian W's tentatively dubbed W4x. There is a single full sequence of a 'pure' W1 (with the 7864 change but no coding region changes leading to W1a, b, etc) which lacks 16292T. If noncoding-region changes are taken into account, this would be the pure W1c, ancestral to W1c1 and W1c2.

Loss of 16223

This marker was inherited from N and seems quite stable. Its disappearance occurs only 1% of the time, in W1's and W3's.

119C

This HVR2 change is noted in the single example of W6. Otherwise all of the examples are W1's. A single complete sequence shows it to be associated with a coding region change at 9695. It is an interesting type in that it seems (so far) to be completely associated with the Atlantic Coast of Western Europe.

Far Eastern W's

There is a very distinctive group of W's so far identified in the Far East - from Tibet and China to Thailand. These have a unique 16111T 16172C 16189C 16223T (no 16292T) 16362C motif. So far none have been tested for coding region changes, but they are probably either a type of W4 (based solely on the 16362C marker) or a different major clade.

194T as clade defining

Considering the HVR2 194T change as relevant, an alternate clade nomenclature is possible as illustrated here.

Data Sources

There are now 1212 W results available - 50 full sequences, 14 coding region-only, 469 HVR1 and HVR2, 302 HVR-1-only Genographic results with information on W3, and 427 HVR-1-only results. Some of these are duplicates - the result of the same individual submitted to multiple sources. These 1212 results group into 200 unique HVR1 motifs, and 179 unique HVR1 + HVR2 motifs. Many motifs are related to a certain geographic region of the Old World. See the entries under each clade for details.

The consolidated database went back to original sources and was revalidated and updated for this analysis. These sources were:

• Full sequence W's as collated by Peter Fischer
• FTDNA W Group members as of December 2007
• Mitosearch W sequences as of December 2007
• Sequences for W's in Asia as published by Cordaux et al. 2003 European Journal of Human Genetics (2003) 11, 253–264; and Metspalu et al. BMC Genetics 2004 5:26 doi:10.1186/1471-2156-5-26
• Blood of Isles Project HVR1 sequences for Britain
• Genographic results as corrected and released here
• SMGF sequences as laboriously extracted from the SMGF database by Peter Fischer.