STR Wars: Is diversity meaningful? more meaningful than Hg frequency?

[Jean M]

The Ian Cam are distinguished by a mutation to 10 at 391.  This is common to all the entries to date.  The Clan Gregor has over 600+ entries; Grier, Grieg, Gregory etc.  The Ian Cam are one set of the entries but assert they are descendants of the founder of the Clan Gregor name.  So, if you go to the FtDNA Clan Gregor website and observe the entries, you will that the Ian Cam are a pretty homgeneous group.  2124 is a direct descendant of the Clan founder and has had no observable mutations.

OK. Things are now crystal clear on certain points.  http://www.familytreedna.com/public/macgregor/ says that 2124 is descended from the MacGregors of Glencarnoch (Chief’s line), and that kit fits into the MacGregor (Ian Cam) section of the results. So we can assume that the Ian Cam lot are descendants of the founder of the clan, if by "clan", we mean persons descended from the first chief that historians can identify (according to Wikipedia) Gregor "of the Golden Bridles." Gregor's son, Iain Camm ("of the One-Eye") succeeded as the second Chief sometime prior to 1390.

Did you attempt to find a date for the founder of the Ian Cam section alone? Or did you include a lot of people named McGregor who may have no connection whatsoever with the said line of chiefs no matter how much they wish they did?

[ironroad41]

Our TMRCA  estimate is our best guess when the modal occurred and the time to the SNP defining that modal can't be too far off.  It might not be O'neill of the nine hostages, but it sure was probably a close relative (re M226).

I think that you mean M222 and in fact there is no evidence that Niall of the Nine Hostages carried M222. If he is anything more than fiction, he was from a part of Ireland which is actually low on M222. The whole idea that he was the ancestor of the men in Donegal carrying M222 rested on genealogies which were tampered with c. 700 AD to make this famous person the ancestor of various families of Donegal, who then claimed to be the Northern Uí Néill. See Irish Surnames and y-DNA: Uí Néill

You may be chasing a similar will-o'-the-wisp with the founder of Clan Gregor. I can't really tell from what you have written. In general men of the same surname who are actually related should turn out to have a common ancestor at around the period that surnames developed, and that has been found to be the case for some  that have been investigated. However if you have been looking at everyone with a McGregor surname, including those not known to be related by paper trail to the clan chiefs, that could be throwing you out. As you know, not all McGregors will be descended from the same Gregor. Some may not be descended from a Gregor at all. Though we can expect some not descended from the clan founder to be at least in the same haplogroup, since R1b-L21 is so common in Scotland. (I take it that the chief's line is L21.)  

If I've got it right, the 'Ian Cam' McGregor group is a younger subclade branching off the very large 'Scots Modal' cluster around 600 years ago. 'Scots Modal' is L21+ but I think it has yet to discover its defining SNP downstream of L21.

Regarding the O'Neill/UiNeill, this is what I picked up: One group is the L21+, M222+ (Ui Neill), a very large cluster known as the 'Nial cluster'/'NW Irish' and contains many surnames.

Another cluster, one that I don't think you mention in the link, (with a number of O'Neill and MacShane surnames) is possibly P312* (named 'O'Neill Variety' or O'Neill Variant') and has not yet discovered its defining SNP downstream of P312 (L21- , U152- and Z196- so far).

You are right on re: the clan gregor.  I believe I have shown that the R1b of Clan Donald, MacMillans and Buchanans all converge about 0 to 200 BC to what is known as the Scots Modal.

Of current interest to us is that 2124 the clan chieftain is having his entire genome evaluated, along with 6 other selected scots, by Jim Wilson.  This might help identify some new SNP's in his line leading to the scots modal?

[Mikewww]

.. The Ian Cam are destringuished by a mutation to 10 at 391.  ..

One STR, even a slow one, is not enough to discern with any certainty.

..So, if you go to the FtDNA Clan Gregor website and observe the entries, you will that the Ian Cam are a pretty homgeneous group.  2124 is a direct descendant of the Clan founder and has had no observable mutations.

How many generations back does 2124's pedigree go?

You are right on re: the clan gregor.  I believe I have shown that the R1b of Clan Donald, MacMillans and Buchanans all converge about 0 to 200 BC to what is known as the Scots Modal.

I think I lost something in this discussion. You've mentioned Z253+ before. Are you saying the Ian Cam descendants have both Z253+ and Scots Modal people AND there is a single Ian Cam founding lineage?

EDIT: It looks like you are saying they do have separate founding lineages. If so, I don't think it is useful to determine the TMRCA for this mixed up group. It would be more appropriate to find all of the subclades within the group, be they Z253+, DF21+, M222+ or whatever and do the TMRCA for that level of the Y DNA tree involved. If Z253, DF21 and M222 covered it then either L21 or DF13 would be the lowest level of the Y DNA tree where the whole group of Ian Cam comes together. If that is the case, there is no use calculating the TMRCA for Ian Cam. Just calculate the TMRCA for all of L21 or all of DF13 and that would as close as we could probably get to a common ancestor for all of Ian Cam.   There people outside of Ian Cam, then that are more closely related to Ian Cam members than all Ian Cam members are to themselves.

Can you ask the project administrator to turn on the Y DNA SNP report?

[Jean M]

The surname McGregor actually occurs before Ian Cam. Or at least someone called the son of Gregor. Duncan M'Greghere is recorded in 1292.

Where people may be getting confused is thinking that earlier McGregors are something to do with the later chiefs. Anyone could name a son Gregor.

[ironroad41]

The Ian Cam are distinguished by a mutation to 10 at 391.  This is common to all the entries to date.  The Clan Gregor has over 600+ entries; Grier, Grieg, Gregory etc.  The Ian Cam are one set of the entries but assert they are descendants of the founder of the Clan Gregor name.  So, if you go to the FtDNA Clan Gregor website and observe the entries, you will that the Ian Cam are a pretty homgeneous group.  2124 is a direct descendant of the Clan founder and has had no observable mutations.

OK. Things are now crystal clear on certain points.  http://www.familytreedna.com/public/macgregor/ says that 2124 is descended from the MacGregors of Glencarnoch (Chief’s line), and that kit fits into the MacGregor (Ian Cam) section of the results. So we can assume that the Ian Cam lot are descendants of the founder of the clan, if by "clan", we mean persons descended from the first chief that historians can identify (according to Wikipedia) Gregor "of the Golden Bridles." Gregor's son, Iain Camm ("of the One-Eye") succeeded as the second Chief sometime prior to 1390.

Did you attempt to find a date for the founder of the Ian Cam section alone? Or did you include a lot of people named McGregor who may have no connection whatsoever with the said line of chiefs no matter how much they wish they did?

 I just used a subset of the Ian Cam only.  I have run extensive TMRCA's on sets of other entries and have gotten convergences back into the BC range.    My estimate required that I identify shared mutations such as with the Stirling entries (who are all Ian Cam); further there is an occasional multistep and I had to exclude the entry with a mutation at 426, since he, drove the estimate back more than 200 years.  With around 39 entries and the caveats above, I get 1350 +/-100 as the best estimate for the occurrence of the 391 mutation?  There is a lot that can be learned in studying one set of data in detail.  

I should also mention that I did the same analysis with the smaller Kerchner family and got 1650 as the TMRCA.

I don't want to bring up old "discussions", but I asked Ken and VV repeatedly to use the variance approach on these two sets of data and they  never responded.

[Jean M]

 I just used a subset of the Ian Cam only. ..    My estimate required that I identify shared mutations such as with the Stirling entries (who are all Ian Cam); further there is an occasional multistep and I had to exclude the entry with a mutation at 426, since he, drove the estimate back more than 200 years.  With around 39 entries and the caveats above, I get 1350 +/-100 as the best estimate for the occurrence of the 391 mutation..

So that would be about 660 AD. So your feeling is that the mutation occurred a lot earlier than the 14th century Gregor "of the Golden Bridles". Makes sense. So where do we go from there? You feel that people in the Ian Cam section are not necessarily descended from Gregor "of the Golden Bridles", but from some ancestor of his. I see your point. I was wrong to assume that they were all descended from the founder. I freely confess it! :)

[Jdean]

I can see Ken continues to work on enhancements.  I currently use his Generations7 methodology in conjunction with the Haplotype_Data spreadsheets I maintain for R1b deep clade tested people. I'll let initial testing settle a little and then I'll incorporate 111T version.

Now is the time to consider upgrading to 111 STRs if you haven't already.

As far as I can tell the major change (apart from the obvious additional loci) is the mutation rates, I wonder where they came from ?

Rather confusingly there are two sets, the ones Ken is using are in the row named 'backup'.

[Jdean]

Of current interest to us is that 2124 the clan chieftain is having his entire genome evaluated, along with 6 other selected scots, by Jim Wilson.  This might help identify some new SNP's in his line leading to the scots modal?

That's really good news, finding a SNP that defined the Scots Modal would be something to look forward to. Any idea if the data will be made public ?

[ironroad41]

Of current interest to us is that 2124 the clan chieftain is having his entire genome evaluated, along with 6 other selected scots, by Jim Wilson.  This might help identify some new SNP's in his line leading to the scots modal?

That's really good news, finding a SNP that defined the Scots Modal would be something to look forward to. Any idea if the data will be made public ?

As I said there are 6 other Scots being included by Jim.  I have no idea when the data will be generated nor who will announce what they find,  It could be quite political since Jim thinks the MacGregors are of Pictish descent as do Woolf and others at Edinborough.

[ironroad41]

.. The Ian Cam are destringuished by a mutation to 10 at 391.  ..

One STR, even a slow one, is not enough to discern with any certainty.

..So, if you go to the FtDNA Clan Gregor website and observe the entries, you will that the Ian Cam are a pretty homgeneous group.  2124 is a direct descendant of the Clan founder and has had no observable mutations.

How many generations back does 2124's pedigree go?

You are right on re: the clan gregor.  I believe I have shown that the R1b of Clan Donald, MacMillans and Buchanans all converge about 0 to 200 BC to what is known as the Scots Modal.

I think I lost something in this discussion. You've mentioned Z253+ before. Are you saying the Ian Cam descendants have both Z253+ and Scots Modal people AND there is a single Ian Cam founding lineage?

EDIT: It looks like you are saying they do have separate founding lineages. If so, I don't think it is useful to determine the TMRCA for this mixed up group. It would be more appropriate to find all of the subclades within the group, be they Z253+, DF21+, M222+ or whatever and do the TMRCA for that level of the Y DNA tree involved. If Z253, DF21 and M222 covered it then either L21 or DF13 would be the lowest level of the Y DNA tree where the whole group of Ian Cam comes together. If that is the case, there is no use calculating the TMRCA for Ian Cam. Just calculate the TMRCA for all of L21 or all of DF13 and that would as close as we could probably get to a common ancestor for all of Ian Cam.   There people outside of Ian Cam, then that are more closely related to Ian Cam members than all Ian Cam members are to themselves.

Can you ask the project administrator to turn on the Y DNA SNP report?

 The Ian Cam have that distinguishing feature, but obviously there is no info in that mutation.  I use some 39 dys loci and a set of entries of about the same size for analysis.  I would dearly love someone besides myself to run a TMRCA estimate on the Ian Cam?

The Ian Cam are Z253-, I and the clan moderator are +.  Until we get the Jim Wilson results on 2124, we have no info on any SNP except for R-L21+ for the clan.  As I said, I took the Ian Cams founders haplotype, a converged MacMillan, a converged R!b Clan Donald set and got a TMRCA c. 0 to 200 BC,  for all we know, they are all just R-L21.  Note Campbells seem to fit in here also.

I'm very confused with your one para.  I can't make sense out of it.  All the data we have is that the Chief is R-L21+ and therefore all the Ian Cam are.  As I said we will learn more (hopefully) after Jim Wilson runs his studies.

[Mikewww]

.... It looks like you are saying they do have separate founding lineages. If so, I don't think it is useful to determine the TMRCA for this mixed up group. It would be more appropriate to find all of the subclades within the group, be they Z253+, DF21+, M222+ or whatever and do the TMRCA for that level of the Y DNA tree involved. If Z253, DF21 and M222 covered it then either L21 or DF13 would be the lowest level of the Y DNA tree where the whole group of Ian Cam comes together. If that is the case, there is no use calculating the TMRCA for Ian Cam. Just calculate the TMRCA for all of L21 or all of DF13 and that would as close as we could probably get to a common ancestor for all of Ian Cam.   There people outside of Ian Cam, then that are more closely related to Ian Cam members than all Ian Cam members are to themselves...

.... I'm very confused with your one para.  I can't make sense out of it.  All the data we have is that the Chief is R-L21+ and therefore all the Ian Cam are.  

I was bit wordy.

All I'm saying is that if you have a mixed bag of subclades in Ian Cam there is no use calculating a TMRCA for it.  

It would be better to calculate the TMRCA for all of the lowest common level of the Y DNA that encompasses all of the Ian Cam members.  In this case, it may be that everyone is some version of L21+.  Some are Z253+ some are Z253- and some maybe L21*.  If that is the case then you might as well just look at the TMRCA for all of L21 as the TMRCA ancestor for Ian Cam.  DYS391 or any other STR commonality are just coincidental.

[Jean M]

It would be better to calculate the TMRCA for all of the lowest common level of the Y DNA that encompasses all of the Ian Cam members.  In this case, it may be that everyone is some version of L21+.  Some are Z253+ some are Z253-

Ironroad is saying that the Ian Cam are all L21+, but Z253-.

Ironroad himself is Z253+, but not in the Ian Cam.

[ironroad41]

.... It looks like you are saying they do have separate founding lineages. If so, I don't think it is useful to determine the TMRCA for this mixed up group. It would be more appropriate to find all of the subclades within the group, be they Z253+, DF21+, M222+ or whatever and do the TMRCA for that level of the Y DNA tree involved. If Z253, DF21 and M222 covered it then either L21 or DF13 would be the lowest level of the Y DNA tree where the whole group of Ian Cam comes together. If that is the case, there is no use calculating the TMRCA for Ian Cam. Just calculate the TMRCA for all of L21 or all of DF13 and that would as close as we could probably get to a common ancestor for all of Ian Cam.   There people outside of Ian Cam, then that are more closely related to Ian Cam members than all Ian Cam members are to themselves...

.... I'm very confused with your one para.  I can't make sense out of it.  All the data we have is that the Chief is R-L21+ and therefore all the Ian Cam are.  

I was bit wordy.

All I'm saying is that if you have a mixed bag of subclades in Ian Cam there is no use calculating a TMRCA for it.  

It would be better to calculate the TMRCA for all of the lowest common level of the Y DNA that encompasses all of the Ian Cam members.  In this case, it may be that everyone is some version of L21+.  Some are Z253+ some are Z253- and some maybe L21*.  If that is the case then you might as well just look at the TMRCA for all of L21 as the TMRCA ancestor for Ian Cam.  DYS391 or any other STR commonality are just coincidental.

Based on a lot of things including family histories many of the Ian Cam can prove from whom their family derives.  By looking at the haplotypes it also becomes evident that this is not a mixed bag.  Theree are (over 70) descended from one man c.1350AD.  As I have been stressing, it has taken a lot of tries and learning to be able to take a set of the Ian Cam, mostly independent in some sense and compute a TMRCA.  You cannot just grab a sample and say voila, you will get the wrong answer more times than not.  But, if you are careful and identify inherited mutations, multisteps etc. it can be done. And I wouldn't call this type of analysis "fudging" the data.  

This is why I have always been skeptical of grabbing a set of data, applying a math rule to it, and then proclaiming veracity.

[Maliclavelli]

They are speaking a lot about this hg. A0 thanks to Ted Kandell and also Argiedude has reappeared on Rootsweb. Ted Kandell, who is what’s more a lovely person, seems a little bit obsessed by this “ancestor”, as if he were searching a new Abraham in the lands of Africa, as though I looked for my to be an European in this 2.8% of Neanderthal. I studied this haplogroup and put some of them on ySearch. About this I tend to agree with Anatole Klyosov, i.e. that this line (I think unique), survived from the prehistory, belongs probably to a hominid whom back migrants from Eurasia mixed with, taking from him their to be Africans.

But to demonstrate (once more) to Mikewww my theories, we could take my haplotype (KV7Y2) R-L150+ and that of Anutechia (MF7MA), A0, separated from our line at least 160,000 years ago, perhaps much more says Anatole Klyosov.

DYS393        12        13
DYS390        24        19
DYS19          15         16
DYS391        10          10
DYS385a      11         15
DYS385b      14         17
DYS426        12         13
DYS388        12         11
DYS439        12          12
DYS389I        13          13
DYS392         12         12
DYS389II       29         30
DYS458          16         16
DYS459a          9           8
DYS459b         10          9
DYS455           11          10
DYS454            11           11
DYS447            24          24
DYS437            15          12
DYS448            19          21
DYS449            29           32.2
DYS464a          14          13
DYS464b          14           15
DYS464c           16          15
DYS464d           17          18
DYS460             10          12
GATAH4.1          11            9
YCAIIa               19            16
YCAIIb               23            18
DYS456              16           15
DYS442              12           12
DYS438              12           16
DYS444              12            12
DYS446              13          19
DYS461              11           12
DYS462              11            13
YGATAA10          14             12
DYS635              23             15
GGAAT1B07        10             10
DYS441              14            13
DYS445              12             12
DYS452               30              31
DYS463               24              20

If I have remembered well my values (but a little changes), we have:

MR: 0,0022
(74:86): 0,0022=391 generations
Your theories (Nordtvedt, Vizachero, Klyosov etc. etc.) are completely absurd. At this level also Zhivotovsky isn’t worth.

[Mikewww]

... But to demonstrate (once more) to Mikewww my theories, we could take my haplotype (KV7Y2) R-L150+ and that of Anutechia (MF7MA), A0, separated from our line at least 160,000 years ago, perhaps much more says Anatole Klyosov......
Your theories (Nordtvedt, Vizachero, Klyosov etc. etc.) are completely absurd. At this level also Zhivotovsky isn’t worth.

As you said, these are the theories of others, not mine. I'm just using what folks like Ken Nordtvedt, Marko Heinila and John Chandler have provided. These folks are true scientists, albeit not in population genetics.... however, they know their math.  I'm just convinced their mathematical models are useful.

... and there you go again, arguing by exception.  The value of statistics is to look at large groups. It is not useful to apply statistics to (I'll use your word,) "absurd" cases.  There is not much use in applying a methodology to two modern people who's SNP phylogeny indicates they are tens  (or hundreds) of years separated. Anutechia (MF7MA) is a modern person. He is not ancient. His lineage has separated from us long, long ago. The branches of the Y DNA family tree grow in many directions. The branches do not just grow away from each other (diverge.)  They also cross (converge.)  What you should be interested in is not your GD from Anutechia but time to your ancestal alleles, that of your common ancestor, a truly ancient person who died probably over 100K or double (?) years ago.  Even so, if you want to estimate such a number, why look at just two people? Why not look at the thousands of long haplotypes available - using the value of statistics?

Oh, speaking of "absurd", is any one saying the linear duration of these STRs lasts for 100's of thousands years?  Your exception argument case should not be used to consider a mathematical model not designed to handle it. Even earlier on this thread, on reply #18, you'll find:

...

... For old haplogroups (e.g. more than 25 ky old) the problem of non-linear accumulation of GD due to marker saturation becomes the dominant problem. Creating trees from STRs in this timeframe is typically not necessary, thankfully, now that our SNP-based trees are so much more complete than they were several years ago.

http://archiver.rootsweb.ancestry.com/th/read/GENEALOGY-DNA/2012-04/1334313410

Not one is saying STR linearity durations aren't an issue over the periods of time in your case. Marko Heinila's analysis, which I've posted on this thread (see reply #146), lists STR durations for most of the common STRs.

....
Nevertheless, some STRs probably do behave non-linearly outside of certain time ranges. Marko Heinila addressed this with a statistical analysis across tens of thousands of haplotypes.
....

Very few STRs might possibly apply to your exception case, but I doubt if Marko would recommend using any of them for 200k year periods. We know Vizachero would not recommend it.

[Mikewww]

... Please don't misinterpret things out of context. There are many disagreements among the mathematicians. However the thoughtful aggregation of STR clocks to estimate the relative age of clades is useful, no doubt.  I expect to see new breakthroughs over the next couple of years... maybe JeanL has one for us.  I think Heinila developed some new forms of analysis.

I can see Ken continues to work on enhancements.  I currently use his Generations7 methodology in conjunction with the Haplotype_Data spreadsheets I maintain for R1b deep clade tested people. I'll let initial testing settle a little and then I'll incorporate 111T version.

I have upgraded my excel program for estimating intra and inter clade variance based age estimates for y haplotypes. Generations111T now takes haplotypes which include all the 111 standard FTDNA STRs (although 11 of the multi-copy ones are not used). But haplotype collections of mixed STR numbers can be used. I like to think the upgrade program is also more user friendly than the Generations7 it replaces.

Generations111T can be downloaded from link below...

http://archiver.rootsweb.ancestry.com/th/read/Y-DNA-HAPLOGROUP-I/2012-05/1335879324

Why not have a play with it ?

It's easy enough to use and you can delete entire loci from the calculation to see how it effects the outcome.

BTW does anybody know where the mutation rates Ken used in Generations111T are published ?

The mutation rates that Ken used in Generations111T are from Marko Heinila; I have so far not been able to find his mutation rates online.
....
I got a note from Marko.  He is still recovering, but he said that his new rates are about 15% faster than Chandlers.  That explains most of  the difference between the 800 years BP for the Ian Cam he had calculated previously and MJosts data? I think he will be on line soon(hope).  

This is interesting. Marko Heinila, as noted earlier in this thread, has evaluated mutation rates by STRs, up and down rates as well as multi-steps.  He actually comes out with faster mutation rates than Chandler's germ line rates.   This puts the situation even more at odds with the very slow evolutionary rates that Zhivotosky et al uses.

[Maliclavelli]

Mike, of course I am not convinced from your answer, also because you take something from VV, and you know how much I esteem him. He tested Romitti for L150 without his permission, after denied the importance of that SNP, and disagreed with FTDNA when they accepted that SNP and abandoned his work of administrator. After came back to his plants, then returned and tried to make me banned again from Dna-forums (that it rests in peace!).
He, like administrator of "ht35 project", had the sample of DeMao whom recognized like R1b1* anly after many letters of mine, he had the sample of Mangino (the Tuscan Mancini) whom it were enough to test for a WTY to undestand where R1b1a2* was born, and he didn't anything for this, he struggled for an Eastern origin of R1b, giving bad suggestions to you all.
Your analyses were all wrong. It becomes to be evident from these R1b found in Germany of 4600 YBP, and which surprise if they tested some SNPs downstream M269! But I am sure, as I was sure of this also some days ago and I wrote this, that they will find in Italy and in Tuscany above all some R1b not of 5000 years old, but at least the double.
Where your analyses will go then?

[JeanL]

This is interesting. Marko Heinila, as noted earlier in this thread, has evaluated mutation rates by STRs, up and down rates as well as multi-steps.  He actually comes out with faster mutation rates than Chandler's germ line rates.   This puts the situation even more at odds with the very slow evolutionary rates that Zhivotosky et al uses.

Mike, the problem here is that the robusticity of these mutation rates are being tested in sets that share common ancestry in relatively short time frame. Then, those values are being extrapolated to time frames that are presumed to be three and four times longer. There is variability in the mutation rate as a function of repeats. In a time frame of 1000-1750 ybp, even the fastest markers would give you at most 2-3 mutations, so the effects of the variability can be reduced by:

a)Analyzing  a large number of markers, because even within the time frame of 1000-1750 ybp there is a possibility that 1 mutation is in fact 3, and that 0 is equal to 2 due to back mutations. However when a large number of markers are used that probability is greatly reduced. That is in fact where the law of large number comes into play.

b)When one calibrates the mutation rate of a haplotype, it will likely yield very good results if there are only 2 or 3 mutation rates per marker that are being averaged, however it doesn’t mean that one could use that average and extrapolate it to an older time frame, because now there are new mutation rates on each microsatellite that are not being accounted for.

Let me explain part b) slightly so that people might better understand. Say: one clan has a common ancestor that live 1000 ybp. You look at  DYS XXX and find that the maximum amount of mutations that have occurred are 2, this means that there were two mutation rates involved in the process(Assuming no back mutations occurred) so when DYS XXX mutated from modal to mut-1 it had mutation rate “a”, and when it mutated from mut-1 to mut-2 it had mutation rate “b”. When one does the calibration method, the mutation rate “c” is actually a very close estimate of the average mutation rate for that time span. Now when you take mutation rate “c” and try to apply it to a longer time span, what happens is that “c” was perfect for the time span in consideration, but in a longer time span one might encounter that there is “different modal” to mut-01 with mutation rate “a0”, and then mut-01 to mut02 with mutation rate “a1”, and so on until you actually reach mutation rates “a” and “b”. For simplicity purposes, I only assumed forward mutations, but in reality in could go anywhere, which further adds more uncertainty to the analyses. The problem is that when mutation rate “c” is applied to time frames that are much longer than where “c” was estimated, one is automatically assuming that all the mutation rates in the new time period being analyzed would average out to “c”, or a number close to it. In reality, due to the variability in the mutation rates, it doesn’t.  This is the reason why I’m thinking of creating a case-control simulation where I can implemented a mutation rate that is a function of the repeat number, and see how good of a fit do calibrated mutation rates found using say 50 generations are for sets that have common ancestry at 100 generations.  This should take me a while, because I want to design a project that could be used for multiple things, and I just got out of school this week, and I’m graduating Sunday. So I want to take a little break, before I get to it.

[Jdean]

This is interesting. Marko Heinila, as noted earlier in this thread, has evaluated mutation rates by STRs, up and down rates as well as multi-steps.  He actually comes out with faster mutation rates than Chandler's germ line rates.   This puts the situation even more at odds with the very slow evolutionary rates that Zhivotosky et al uses.

Mike, the problem here is that the robusticity of these mutation rates are being tested in sets that share common ancestry in relatively short time frame. Then, those values are being extrapolated to time frames that are presumed to be three and four times longer. There is variability in the mutation rate as a function of repeats. In a time frame of 1000-1750 ybp, even the fastest markers would give you at most 2-3 mutations, so the effects of the variability can be reduced by:

a)Analyzing  a large number of markers, because even within the time frame of 1000-1750 ybp there is a possibility that 1 mutation is in fact 3, and that 0 is equal to 2 due to back mutations. However when a large number of markers are used that probability is greatly reduced. That is in fact where the law of large number comes into play.

b)When one calibrates the mutation rate of a haplotype, it will likely yield very good results if there are only 2 or 3 mutation rates per marker that are being averaged, however it doesn’t mean that one could use that average and extrapolate it to an older time frame, because now there are new mutation rates on each microsatellite that are not being accounted for.

Let me explain part b) slightly so that people might better understand. Say: one clan has a common ancestor that live 1000 ybp. You look at  DYS XXX and find that the maximum amount of mutations that have occurred are 2, this means that there were two mutation rates involved in the process(Assuming no back mutations occurred) so when DYS XXX mutated from modal to mut-1 it had mutation rate “a”, and when it mutated from mut-1 to mut-2 it had mutation rate “b”. When one does the calibration method, the mutation rate “c” is actually a very close estimate of the average mutation rate for that time span. Now when you take mutation rate “c” and try to apply it to a longer time span, what happens is that “c” was perfect for the time span in consideration, but in a longer time span one might encounter that there is “different modal” to mut-01 with mutation rate “a0”, and then mut-01 to mut02 with mutation rate “a1”, and so on until you actually reach mutation rates “a” and “b”. For simplicity purposes, I only assumed forward mutations, but in reality in could go anywhere, which further adds more uncertainty to the analyses. The problem is that when mutation rate “c” is applied to time frames that are much longer than where “c” was estimated, one is automatically assuming that all the mutation rates in the new time period being analyzed would average out to “c”, or a number close to it. In reality, due to the variability in the mutation rates, it doesn’t.  This is the reason why I’m thinking of creating a case-control simulation where I can implemented a mutation rate that is a function of the repeat number, and see how good of a fit do calibrated mutation rates found using say 50 generations are for sets that have common ancestry at 100 generations.  This should take me a while, because I want to design a project that could be used for multiple things, and I just got out of school this week, and I’m graduating Sunday. So I want to take a little break, before I get to it.

There is also the issue of how long a generation is, presumably this was estimated and would end up being built into the mutation rates.

[Mikewww]

This is interesting. Marko Heinila, as noted earlier in this thread, has evaluated mutation rates by STRs, up and down rates as well as multi-steps.  He actually comes out with faster mutation rates than Chandler's germ line rates.   This puts the situation even more at odds with the very slow evolutionary rates that Zhivotosky et al uses.

Mike, the problem here is that the robusticity of these mutation rates are being tested in sets that share common ancestry in relatively short time frame. Then, those values are being extrapolated to time frames that are presumed to be three and four times longer. There is variability in the mutation rate as a function of repeats.....

I agree that you have valid concerns.  I don't know what the correct mutation rates are.  The only thing that has me leaning towards the germ-line rates being applicable for the STR durations (as depicted by Heinila) is that if I look at the ages of all of the haplogroups (I, G, E, etc.) in context, the resulting TMRCAs make sense, again for reasonable timeframes.

This should take me a while, because I want to design a project that could be used for multiple things,

Very, good. You could do a lot of good to progress the current models or come up with a new one.

and I just got out of school this week, and I’m graduating Sunday. So I want to take a little break, before I get to it.

Congratulations! That's great.  Make sure to enjoy your accomplishment. This is something that can never be taken away from you.

[JeanL]

I agree that you have valid concerns.  I don't know what the correct mutation rates are.  The only thing that has me leaning towards the germ-line rates being applicable for the STR durations (as depicted by Heinila) is that if I look at the ages of all of the haplogroups (I, G, E, etc.) in context, the resulting TMRCAs make sense, again for reasonable timeframes.

I agree with you, germ-line mutation rates are the way to go forward,  however, there is still a lot more case-control checks that ought to be done to assure that things such as variability in mutation rates, loss of linearity do not affect the outcome.

 

Congratulations! That's great.  Make sure to enjoy your accomplishment. This is something that can never be taken away from you.

Yes I’m pretty excited, although the job market here in the US is kind of bad. I’m graduating with a Bachelors of Science in Mechanical Engineering, and I have specialized(i.e. I have a minor) in Bio-Mechanics, specifically computational bio-dynamics.

[ironroad41]

Mutation rates affect long term estimates the most I believe.  The faster mutators are characterized by fairly good sets of father/son meioses but the slower are very infrequent.

I think using tables of entries is questionable.  There are too many sets of families with redundant mutations. It is blatantly evident in the Ian Cam.  So, you overcount the occurrence of mutations.

As I said on another thread, 388 is a classic example of a dys loci which changes rate with allele value.  The difference in up/down rates would suggest a migration of the modal with time, but I haven't seen that in the data. 

Burgurellas 110 dys loci estimates are subdivided by motif (3, 4, 5 etc.).  Although only 3 and 4 are dominant in their data set.  So the motif appears to affect rate.

There are additional subtleties in which changes in latitude ( migration)  may have a short-term affect.

We are all homo sapiens, but we are also sentient beings and we respond to our environment.  I believe that is what darwin first postulated?

[JeanL]

I found this study while browsing online today, while it is not directly related to humans or Y-STRs for that matter, I think it provides good insights as to what I was talking about regarding the calibration of mutation rates and the time frames.

http://mbe.oxfordjournals.org/content/29/2/707.abstract

The rate of change in DNA is an important parameter for understanding molecular evolution and hence for inferences drawn from studies of phylogeography and phylogenetics. Most rate calibrations for mitochondrial coding regions in marine species have been made from divergence dating for fossils and vicariant events older than 1–2 My and are typically 0.5–2% per lineage per million years. Recently, calibrations made with ancient DNA (aDNA) from younger dates have yielded faster rates, suggesting that estimates of the molecular rate of change depend on the time of calibration, decaying from the instantaneous mutation rate to the phylogenetic substitution rate. aDNA methods for recent calibrations are not available for most marine taxa so instead we use radiometric dates for sea-level rise onto the Sunda Shelf following the Last Glacial Maximum (starting ∼18,000 years ago), which led to massive population expansions for marine species. Instead of divergence dating, we use a two-epoch coalescent model of logistic population growth preceded by a constant population size to infer a time in mutational units for the beginning of these expansion events. This model compares favorably to simpler coalescent models of constant population size, and exponential or logistic growth, and is far more precise than estimates from the mismatch distribution. Mean rates estimated with this method for mitochondrial coding genes in three invertebrate species are elevated in comparison to older calibration points (2.3–6.6% per lineage per million years), lending additional support to the hypothesis of calibration time dependency for molecular rates.

[Mikewww]

I found this study while browsing online today, while it is not directly related to humans or Y-STRs for that matter, I think it provides good insights as to what I was talking about regarding the calibration of mutation rates and the time frames.

http://mbe.oxfordjournals.org/content/29/2/707.abstract

The rate of change in DNA is an important parameter for understanding molecular evolution and hence for inferences drawn from studies of phylogeography and phylogenetics. Most rate calibrations for mitochondrial coding regions in marine species have been made from divergence dating for fossils and vicariant events older than 1–2 My and are typically 0.5–2% per lineage per million years. Recently, calibrations made with ancient DNA (aDNA) from younger dates have yielded faster rates, suggesting that estimates of the molecular rate of change depend on the time of calibration, decaying from the instantaneous mutation rate to the phylogenetic substitution rate. aDNA methods for recent calibrations are not available for most marine taxa so instead we use radiometric dates for sea-level rise onto the Sunda Shelf following the Last Glacial Maximum (starting ∼18,000 years ago), which led to massive population expansions for marine species. Instead of divergence dating, we use a two-epoch coalescent model of logistic population growth preceded by a constant population size to infer a time in mutational units for the beginning of these expansion events. This model compares favorably to simpler coalescent models of constant population size, and exponential or logistic growth, and is far more precise than estimates from the mismatch distribution. Mean rates estimated with this method for mitochondrial coding genes in three invertebrate species are elevated in comparison to older calibration points (2.3–6.6% per lineage per million years), lending additional support to the hypothesis of calibration time dependency for molecular rates.

I haven't read their paper but I'm not really following the whole line of reasoning.
Is this still applicable given the long period of time?  I may misunderstand but they mention a "two epoch" model.  An Epoch is a long, long time.. way behind our period of interest.

In a geologic time frame, Epochs divide Periods into smaller chunks, and the lengths of Epochs range in the tens of millions of years. In the most recent Era, the Cenozoic, they have ranged from 2 million to 22 million years long. There have been seven Epochs in the last 70 million years or so, and the current Epoch is called the Holocene. You can see the Geologic Time Scale

http://answers.ask.com/Science/Nature/how_long_is_an_epoch

[JeanL]

This is on the STR Wars thread as well, I compare variance both with a set of mixed speed markers and then with a subset that meet a 7000 year linear duration according to Marko Heinila. In all cases multi-copy (i.e. CDY, 464, 459, etc.) and null potential STRs (i.e.439, 425) are thrown out. In other words, in variance comparisons I have a rational, quality driven approach to using STRs. As far as Ken's TMRCA tool, he is essentially using everything but multi-copy STRs and requires the user to adjust nulls to an incremental GD.

Indeed you did:

http://www.worldfamilies.net/forum/index.php?topic=10513.msg129272#msg129272

However:

1) I believe that even though you used the 36 most linear(Having linearity longer than 7000 ybp) STRs you used per Marko.H calculations still show a wide variability in mutation rates. In fact, only 24 STRs in the 111 marker set provided by Marko.H have mutation rates lower than 10^-3, thus making them slow STRs.

2)The calculations were perfomed on the R1b-L21+ dataset from the FTDNA Projects which are heavily populated by folks of British descent, so if the TMRCA of L21 in Britain is indeed 4000 ybp, then both most linear or mixed sets of STR ought to give you the same result.


3)In a nutshell you can argue that based on the calculations on R1b-L21+ the difference between using 36 STRs that have a linearity of 7000+ybp vs. a mixed bag of 49 appears to have little effect on variance for a set that is mostly populated by British guys.

4)Now can you extrapolate those conclusions to say P312+(i.e. DF27, U152,etc) folks from elsewhere in Europe? I for once wouldn’t do it.