World Families Forums - STR Wars: Is diversity meaningful? more meaningful than Hg frequency?

Welcome, Guest. Please login or register.
December 21, 2014, 10:20:48 PM
Home Help Search Login Register

+  World Families Forums
|-+  General Forums - Note: You must Be Logged In to post. Anyone can browse.
| |-+  R1b General (Moderator: rms2)
| | |-+  STR Wars: Is diversity meaningful? more meaningful than Hg frequency?
« previous next »
Pages: 1 ... 6 7 [8] 9 10 ... 14 Go Down Print
Author Topic: STR Wars: Is diversity meaningful? more meaningful than Hg frequency?  (Read 27745 times)
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #175 on: April 28, 2012, 11:39:11 AM »

... But, I do believe that the current estimates of time for many of our R-1b subclades is too short?  I am asking Why?

Why do you think the estimates from Nordtvedt and Heinila are low? You've listed concerns about up/down mutation rates, multi-steps, etc.  That is fine, but the concerns do not necessarily make the estimates top young. They could be too old. I have no reason to think the estimates and their confidence ranges are automatically always too young.  When do you believe R1b entered or was born in Europe?

Note: it isn''t only my opinion here, its about half or more of the genetic "guru" community.  So, I am not Don Quixote out chasing windmills.  Reasonable minds have made the same observation.  I think the best chance to resolve the issue is "data mining", which I have some hands-on experience in.

I want to be clear that I agree mutation rates are an open issue. This is why I generally only calculate STR variance myself and only use other peoples' (i.e. Nordtvedt's or Heinila's) TMRCA estimates.

Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #176 on: April 28, 2012, 11:43:51 AM »

... So, I am not Don Quixote out chasing windmills....

I've never brought up Don Quixote, but since you've brought it up it is interesting that we have folks they have "slayed sacred monsters" - or something along that line.
« Last Edit: April 28, 2012, 11:44:12 AM by Mikewww » Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
JeanL
Old Hand
****
Offline Offline

Posts: 425


« Reply #177 on: April 28, 2012, 12:12:46 PM »


Nothing is every ideal, which is what science tries to deal with. Everything I've read from Nordtvedt is that he is evaluating the data and trying to account for what he calls "fictions." That doesn't mean the mathematical models don't work.. it's a matter of precision.

Well tell me something I don’t already know. Ok, I don’t mean to be harsh, but yeah, the main problem here is the great number of assumptions that come into play on the “data” that Dr.Nordtvedt is evaluating. Like I said before, when we look at father-son’s pairs, we know what the allele at a given loci is, we know in that sense what the modal(ancestral) is, but even in that scenario there  could be cases of somatic mutations. As you can see in the study I provided, which used 23 autosomal STRs, which actually are a lot more stable(Which is self-evident in the case that even at the autosomal level, the paternal germ-line is three folds the maternal germline) than Y-chromosome STRs, there were 31 losses vs.22 gains. If one was to run a mathematical simulation across a generation you will never get a disparity of 9 mutations, at best 1-2. So again, this is family pairs, where all factors are known, and with a good precision we can determine mutation rates.

When it comes to the X Clan which has 100 participants, and a presumed common ancestor who lived 1300 ybp, the ancestral haplotype of that ancestor isn’t known, whether everyone in that generation is equally far removed, or at least on average far removed x generations from that ancestor isn’t know, the mutation rate is assumed to be constant, or to change very little, in order to estimate mutation rates in a time span of 1300 ybp. Often times the mutation rate will remain constant, because even with the fastest STRs the most mutations that will happen is 2 or 3 in 1300 ybp. However when one gets into 3000, 4000, 5000 ybp those fast mutating STRs probably incremented their mutation rate quote significantly along the process and one is presuming that 2*10^-3 measured for loci DYS-B from Clan X would do it, but in reality the average mutation rate could be something like 0.5*10^-3 for loci DYS-B when the time span is 3000-5000 ybp. 

I have no issue with your concern. I don't think your concern has a significant impact given the number of markers I'm evaluating. As I've shown on this thread, the results seem to show a correlation with the known SNP defined phylogenetic Y DNA tree, at least for the R1b subclade.  We know Ken thinks the correlation of STR diversity to time applies to the Hg I subclades appropriately as you can see from his web site.


... but I am NOT an expert so I'll request again:

Please, please, please make your challenges directly to Nordtvedt on the Rootsweb haplogroup I forum. I'm trying to be a proxy for some of your concerns, but you are much better at arguing your point of view than I am so please go ask Ken.

Of course you don’t think my concern will impact your work, of course none of the folks who are averaging mutation rates, and ignoring changes(loss of linearity) in mutation rate think it will have any impact in their work. If they did think so, then they wouldn’t do it, I don’t think anyone is going to waste a lot of precious time in something they know is wrong.  As for directing my challenges, I don’t want to challenge Dr.Nordtvedt, on the contrary, I prefer to run my own experiments on a case-control scenario, and see what results they yield.
« Last Edit: April 28, 2012, 12:16:01 PM by JeanL » Logged
Maliclavelli
Guru
*****
Offline Offline

Posts: 2158


« Reply #178 on: April 28, 2012, 12:28:07 PM »

I'm not well versed in non-R1b SNPs so please provide as much relevant background as you can on the other haplogroups if that is your area of challenge.
Amongst the thousands of letters I wrote, I have found this about the hg. G2a found in France on a site of 7000 years ago. What is my thinking is said here by Oh-Willeke and Argiedude, so that you doesn’t think I am alone:

Looking closely at the median joining network, the case for the Northern Caucasus as the source of G2a almost anywhere else seems remote. Instead, the Northern Caucasus looks like a recipient of G2a from Mediterranean or the Middle East, and honestly, the Western Mediterranean looks like the best fit for a Northern Caucasus G2a source based on that network with the original G2a bearing men in the Northern Caucasus in this scenario probably arriving by sea, rather than overland from the Middle East, and the migration probably taking place at some time before R1b became common in the Mediterranean, but probably post-Neolithic revolution.

The network is also suggestive of the idea that the Treilles group may have had immediate antecedent in Italy and only more remote antecedents in Iberia.

This direction of migration is quite unexpected (Oh-Willeke).

All TMRCA estimates always produce ages of just 2000 to 5000 years. So apparently the world can trace its y-dna to just 2 or 3 dozen men that lived barely 3000 years ago. It's a little more likely that the entire theory of TMRCA is a piece of ssss (must I mean that all these spreadsheets are spreadshits?) And this study has found a sample that wasn't supposed to exist 5000 years ago, hell, not even 2500 years ago (Argiedude).

Very good you both! Some years ago, after having read a paper of Yunusbaev, I said that the centre of hg. G was the Adriatic, and about the TMRCA I have written a lot on the mutations around the modal laughing of some know-all like Nordtvedt and Klyosov.
Very good you both and a third: me.

P.S. Watching the paper of Yunusbaev, the center of diffusion of hg. G (probably G2a) was Sardinia.

    Re: Neolithic Farmers and the Spread of Indo-European, The Case for Euphratic, etc. « Reply #69 on: June 04, 2011, 04:34:44 AM »

Logged

Maliclavelli


YDNA: R-S12460


MtDNA: K1a1b1e

Maliclavelli
Guru
*****
Offline Offline

Posts: 2158


« Reply #179 on: April 28, 2012, 12:31:52 PM »

I've never brought up Don Quixote, but since you've brought it up it is interesting that we have folks they have "slayed sacred monsters" - or something along that line.
Some sacred monster of mtDNA has been slayed also recently in this forum, and not by me only.
Logged

Maliclavelli


YDNA: R-S12460


MtDNA: K1a1b1e

ironroad41
Old Hand
****
Offline Offline

Posts: 219


« Reply #180 on: April 28, 2012, 12:51:39 PM »


Nothing is every ideal, which is what science tries to deal with. Everything I've read from Nordtvedt is that he is evaluating the data and trying to account for what he calls "fictions." That doesn't mean the mathematical models don't work.. it's a matter of precision.

Well tell me something I don’t already know. Ok, I don’t mean to be harsh, but yeah, the main problem here is the great number of assumptions that come into play on the “data” that Dr.Nordtvedt is evaluating. Like I said before, when we look at father-son’s pairs, we know what the allele at a given loci is, we know in that sense what the modal(ancestral) is, but even in that scenario there  could be cases of somatic mutations. As you can see in the study I provided, which used 23 autosomal STRs, which actually are a lot more stable(Which is self-evident in the case that even at the autosomal level, the paternal germ-line is three folds the maternal germline) than Y-chromosome STRs, there were 31 losses vs.22 gains. If one was to run a mathematical simulation across a generation you will never get a disparity of 9 mutations, at best 1-2. So again, this is family pairs, where all factors are known, and with a good precision we can determine mutation rates.

When it comes to the X Clan which has 100 participants, and a presumed common ancestor who lived 1300 ybp, the ancestral haplotype of that ancestor isn’t known, whether everyone in that generation is equally far removed, or at least on average far removed x generations from that ancestor isn’t know, the mutation rate is assumed to be constant, or to change very little, in order to estimate mutation rates in a time span of 1300 ybp. Often times the mutation rate will remain constant, because even with the fastest STRs the most mutations that will happen is 2 or 3 in 1300 ybp. However when one gets into 3000, 4000, 5000 ybp those fast mutating STRs probably incremented their mutation rate quote significantly along the process and one is presuming that 2*10^-3 measured for loci DYS-B from Clan X would do it, but in reality the average mutation rate could be something like 0.5*10^-3 for loci DYS-B when the time span is 3000-5000 ybp. 

I have no issue with your concern. I don't think your concern has a significant impact given the number of markers I'm evaluating. As I've shown on this thread, the results seem to show a correlation with the known SNP defined phylogenetic Y DNA tree, at least for the R1b subclade.  We know Ken thinks the correlation of STR diversity to time applies to the Hg I subclades appropriately as you can see from his web site.


... but I am NOT an expert so I'll request again:

Please, please, please make your challenges directly to Nordtvedt on the Rootsweb haplogroup I forum. I'm trying to be a proxy for some of your concerns, but you are much better at arguing your point of view than I am so please go ask Ken.

Of course you don’t think my concern will impact your work, of course none of the folks who are averaging mutation rates, and ignoring changes(loss of linearity) in mutation rate think it will have any impact in their work. If they did think so, then they wouldn’t do it, I don’t think anyone is going to waste a lot of precious time in something they know is wrong.  As for directing my challenges, I don’t want to challenge Dr.Nordtvedt, on the contrary, I prefer to run my own experiments on a case-control scenario, and see what results they yield.


I think that your comments are correct and timely.  As I said on my thread I will try to present data that shows that a large number of STR's are bounded in their range, which implies that hidden mutations are prevalent.  I think that hidden mutations will produce the same observable effect you cite, an apparent decrease in mutation rate over longer time spans.  I know the Maori and Gypsy data Zhiv used was not more than a thousand or so years, but if he was using the Faster mutators such as CDYa,b; then I would expect a similar observation would be made and apparently thats what he did?
Logged
Jdean
Old Hand
****
Offline Offline

Posts: 678


« Reply #181 on: April 28, 2012, 03:19:10 PM »

I think that your comments are correct and timely.  As I said on my thread I will try to present data that shows that a large number of STR's are bounded in their range, which implies that hidden mutations are prevalent.  I think that hidden mutations will produce the same observable effect you cite, an apparent decrease in mutation rate over longer time spans.  I know the Maori and Gypsy data Zhiv used was not more than a thousand or so years, but if he was using the Faster mutators such as CDYa,b; then I would expect a similar observation would be made and apparently thats what he did?

I think it's at least reasonably clear that loci don't mutate equally up or down also it's likely that loci with different values will have slightly (probably very slightly) diffrent mutation rates.

However how big is this issue and do we need to worry about it that much, especially when talking about L11 and its offspring ?

DYS492 offers a convenient way to see if loci are likely to behave in a dramatically different way after a mutation.

looking at all the values for this loci in the P312 project I found

9         0.10%
10         0.21%
11         0.94%
12         95.60%
13         1.68%
14         1.47%

where as in the U106 project we see

10         0.12%
12         1.06%
13         95.89%
14         2.82%
15         0.12%


Both sets of data look remarkably similar apart from the obvious fact that 12 is modal in the first set and 13 in the second.

There definitely doesn’t seem to be any particular tendency for U106 to try and get back to a value of 12 and I think it's reasonable to conclude that if somebody who was P312 had a value of 13 at this loci that it would behave in exactly the same way as somebody who had 13 and was U106.
Logged

Y-DNA R-DF49*
MtDNA J1c2e
Kit No. 117897
Ysearch 3BMC9

Mark Jost
Old Hand
****
Offline Offline

Posts: 707


« Reply #182 on: April 28, 2012, 03:57:17 PM »

I added up all the Ups and downs in MikeW's L21 ALL 67 marker spreadsheet by STR and totaled.

More or less from Modal
Allele values higher than the modal: 41789 57%
Allele value lower than Modal: 31974 43%
Out of 67 markers there were 49 STRs had a majority up and 18 down. (73 vs 27%)


Logged

148326
Pos: Z245 L459 L21 DF13**
Neg: DF23 L513 L96 L144 Z255 Z253 DF21 DF41 (Z254 P66 P314.2 M37 M222  L563 L526 L226 L195 L193 L192.1 L159.2 L130 DF63 DF5 DF49)
WTYNeg: L555 L371 (L9/L10 L370 L302/L319.1 L554 L564 L577 P69 L626 L627 L643 L679)
Mark Jost
Old Hand
****
Offline Offline

Posts: 707


« Reply #183 on: April 28, 2012, 04:00:41 PM »

MikeW,

Have you ever tried to incorporate Chandlers mutation rate calculation formula into your spreadsheets to see there was differences in various Haplogroups or Subclades?

MJost
Logged

148326
Pos: Z245 L459 L21 DF13**
Neg: DF23 L513 L96 L144 Z255 Z253 DF21 DF41 (Z254 P66 P314.2 M37 M222  L563 L526 L226 L195 L193 L192.1 L159.2 L130 DF63 DF5 DF49)
WTYNeg: L555 L371 (L9/L10 L370 L302/L319.1 L554 L564 L577 P69 L626 L627 L643 L679)
ironroad41
Old Hand
****
Offline Offline

Posts: 219


« Reply #184 on: April 28, 2012, 04:06:18 PM »

I think that your comments are correct and timely.  As I said on my thread I will try to present data that shows that a large number of STR's are bounded in their range, which implies that hidden mutations are prevalent.  I think that hidden mutations will produce the same observable effect you cite, an apparent decrease in mutation rate over longer time spans.  I know the Maori and Gypsy data Zhiv used was not more than a thousand or so years, but if he was using the Faster mutators such as CDYa,b; then I would expect a similar observation would be made and apparently thats what he did?

I think it's at least reasonably clear that loci don't mutate equally up or down also it's likely that loci with different values will have slightly (probably very slightly) diffrent mutation rates.

However how big is this issue and do we need to worry about it that much, especially when talking about L11 and its offspring ?

DYS492 offers a convenient way to see if loci are likely to behave in a dramatically different way after a mutation.

looking at all the values for this loci in the P312 project I found

9         0.10%
10         0.21%
11         0.94%
12         95.60%
13         1.68%
14         1.47%

where as in the U106 project we see

10         0.12%
12         1.06%
13         95.89%
14         2.82%
15         0.12%


Both sets of data look remarkably similar apart from the obvious fact that 12 is modal in the first set and 13 in the second.

There definitely doesn’t seem to be any particular tendency for U106 to try and get back to a value of 12 and I think it's reasonable to conclude that if somebody who was P312 had a value of 13 at this loci that it would behave in exactly the same way as somebody who had 13 and was U106.

 The distribution about the modal is close in both sets of data.  Whether 5% of the mutations are multisteps is hard to discern.  Look at the behavior of 388 in R1b, I and J.  I think you'll be surprised?

An additional comment would be that this is a very slow mutator and equal values for 13 and 14 in the 312 data set is interesting.  In general, there aren't a lot mutations at this loci.  Two options might be: a. a mutation from 12 to 13 and then from 13 to 14.  Once at 14, there may be descendants who carry that unusual value, or b. a multistep with the same scenario as above.  i.e.  we don't know if the proliferation of 14 is a family event or random event?  I've seen data where, an apparent multistep occurred of about 4 steps and then a population built up around that value.  In the case of 388, we observe a modal 12 the E Hgs, and then successively higher for G(50%  12 and 50% 13),I (14) and J(15) and then a return to 12 for R1a and R1b, as if they had evolved directly E3a and E3b. This data is obtained from the dataset I referenced earlier. Note the width of the distribution for E3a and b is essentially 3, only two states for G, 4 for I and 6 for J.  R1a is unusual in that the modal is 12 (.83) and the allele state 10 has .15.  Clearly a multistep occurred and then a population growth occurred?
« Last Edit: April 28, 2012, 04:23:16 PM by ironroad41 » Logged
Jdean
Old Hand
****
Offline Offline

Posts: 678


« Reply #185 on: April 28, 2012, 04:26:29 PM »

I think that your comments are correct and timely.  As I said on my thread I will try to present data that shows that a large number of STR's are bounded in their range, which implies that hidden mutations are prevalent.  I think that hidden mutations will produce the same observable effect you cite, an apparent decrease in mutation rate over longer time spans.  I know the Maori and Gypsy data Zhiv used was not more than a thousand or so years, but if he was using the Faster mutators such as CDYa,b; then I would expect a similar observation would be made and apparently thats what he did?

I think it's at least reasonably clear that loci don't mutate equally up or down also it's likely that loci with different values will have slightly (probably very slightly) diffrent mutation rates.

However how big is this issue and do we need to worry about it that much, especially when talking about L11 and its offspring ?

DYS492 offers a convenient way to see if loci are likely to behave in a dramatically different way after a mutation.

looking at all the values for this loci in the P312 project I found

9         0.10%
10         0.21%
11         0.94%
12         95.60%
13         1.68%
14         1.47%

where as in the U106 project we see

10         0.12%
12         1.06%
13         95.89%
14         2.82%
15         0.12%


Both sets of data look remarkably similar apart from the obvious fact that 12 is modal in the first set and 13 in the second.

There definitely doesn’t seem to be any particular tendency for U106 to try and get back to a value of 12 and I think it's reasonable to conclude that if somebody who was P312 had a value of 13 at this loci that it would behave in exactly the same way as somebody who had 13 and was U106.

  The distribution about the modal is close in both sets of data.  Whether 5% of the mutations are multisteps is hard to discern.  Look at the behavior of 388 in R1b, I and J.  I think you'll be surprised?

Exactly but the modal is different in both cases, you seem to be suggesting (I think) that there is something special about the modal value itself that causes these distribution patterns but it's just normal distribution.

I will have a closer look at DYS388 but I don't think I will be terribly surprised, I just had a look at a small surname project that I knew had tight groups in J2b, U106, L21, I1 and E1b1b1. They had different values for that loci which isn't a great surprise considering the time to the common ancestor.
Logged

Y-DNA R-DF49*
MtDNA J1c2e
Kit No. 117897
Ysearch 3BMC9

Maliclavelli
Guru
*****
Offline Offline

Posts: 2158


« Reply #186 on: April 28, 2012, 04:30:41 PM »

DYS492 offers a convenient way to see if loci are likely to behave in a dramatically different way after a mutation.
Looking at all the values for this loci in the P312 project I found

9         0.10%
10         0.21%
11         0.94%
12         95.60%
13         1.68%
14         1.47%

where as in the U106 project we see

10         0.12%
12         1.06%
13         95.89%
14         2.82%
15         0.12%


Both sets of data look remarkably similar apart from the obvious fact that 12 is modal in the first set and 13 in the second.
There definitely doesn’t seem to be any particular tendency for U106 to try and get back to a value of 12 and I think it's reasonable to conclude that if somebody who was P312 had a value of 13 at this loci that it would behave in exactly the same way as somebody who had 13 and was U106.
Actually these data are very interesting. They demonstrate that probably R-P312 is a little bit older that R-U106, that, casually, the mutation U106 happened in a man with DYS492=13, and, as the DYS492=13 were very few amongst the ancestor R-L11, this mutation happened amongst those few rather than the great majority with DYS492=12. It is also clear that from then the mutations forwards happen with a slower mutations rate because they start from 13 and not 12, or, better, the mutation from 13 to 14 happens with a faster mutation rate, but from 14 to 15 the mutation is much slower. The mutations forwards are faster than those downwards, but the values of 9 for R-P312 and 10 and 15 for R-U106 are next to the saturation.
Unfortunately we haven’t the data of SMGF about this marker. The mutation rate of DYS492 (Ballantyne) is very slow (3.92x10-4), then the lost of about 5% of the modal may also mean many thousands of years.
Logged

Maliclavelli


YDNA: R-S12460


MtDNA: K1a1b1e

Jdean
Old Hand
****
Offline Offline

Posts: 678


« Reply #187 on: April 28, 2012, 04:35:11 PM »

DYS492 offers a convenient way to see if loci are likely to behave in a dramatically different way after a mutation.
Looking at all the values for this loci in the P312 project I found

9         0.10%
10         0.21%
11         0.94%
12         95.60%
13         1.68%
14         1.47%

where as in the U106 project we see

10         0.12%
12         1.06%
13         95.89%
14         2.82%
15         0.12%


Both sets of data look remarkably similar apart from the obvious fact that 12 is modal in the first set and 13 in the second.
There definitely doesn’t seem to be any particular tendency for U106 to try and get back to a value of 12 and I think it's reasonable to conclude that if somebody who was P312 had a value of 13 at this loci that it would behave in exactly the same way as somebody who had 13 and was U106.
Actually these data are very interesting. They demonstrate that probably R-P312 is a little bit older that R-U106, that, casually, the mutation U106 happened in a man with DYS492=13, and, as the DYS492=13 were very few amongst the ancestor R-L11, this mutation happened amongst those few rather than the great majority with DYS492=12. It is also clear that from then the mutations forwards happen with a slower mutations rate because they start from 13 and not 12, or, better, the mutation from 13 to 14 happens with a faster mutation rate, but from 14 to 15 the mutation is much slower. The mutations forwards are faster than those downwards, but the values of 9 for R-P312 and 10 and 15 for R-U106 are next to the saturation.
Unfortunately we haven’t the data of SMGF about this marker. The mutation rate of DYS492 (Ballantyne) is very slow (3.92x10-4), then the lost of about 5% of the modal may also mean many thousands of years.


Funnily enough I agree with most of what you've written here, which doesn’t happen very often.

You can't say that U106 is younger than P312 from this data alone though, however I thing it probably is but not by much.
Logged

Y-DNA R-DF49*
MtDNA J1c2e
Kit No. 117897
Ysearch 3BMC9

ironroad41
Old Hand
****
Offline Offline

Posts: 219


« Reply #188 on: April 28, 2012, 06:22:33 PM »

In general, I don't believe the modal changes over time for a Hg.  So, I'm not emphasizing the modal value.  Each distribution about the modal appears to be unique.  In part, because of multistep occurrences.  It would be great if the board had someone with a chemical kinetics background who could help enlighten us.

My whole thrust is to get us interested in looking at the data which has been generated and try to understand it.  The model has to fit the data and not the other way around.  One problem with physicists in general is that they love to postulate a theory, the less the data the better, and then prove that it is correct through modelling.   My approach is to use the data to suggest the best model.  FWIW I have an MS in physics.
« Last Edit: April 28, 2012, 06:23:29 PM by ironroad41 » Logged
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #189 on: April 28, 2012, 11:04:36 PM »

MikeW,

Have you ever tried to incorporate Chandlers mutation rate calculation formula into your spreadsheets to see there was differences in various Haplogroups or Subclades?

MJost

I don't apply mutation rates to any formulas in the Haplotype_Data spreadsheets i post. For any time calculation, I use whatever Ken Nordtvedt has in the Y DNA Tools Generations 7 methodology.
Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #190 on: April 28, 2012, 11:07:41 PM »

DYS492 offers a convenient way to see if loci are likely to behave in a dramatically different way after a mutation.
Looking at all the values for this loci in the P312 project I found

9         0.10%
10         0.21%
11         0.94%
12         95.60%
13         1.68%
14         1.47%

where as in the U106 project we see

10         0.12%
12         1.06%
13         95.89%
14         2.82%
15         0.12%


Both sets of data look remarkably similar apart from the obvious fact that 12 is modal in the first set and 13 in the second.
There definitely doesn’t seem to be any particular tendency for U106 to try and get back to a value of 12 and I think it's reasonable to conclude that if somebody who was P312 had a value of 13 at this loci that it would behave in exactly the same way as somebody who had 13 and was U106.
Actually these data are very interesting. They demonstrate that probably R-P312 is a little bit older that R-U106, that, casually, the mutation U106 happened in a man with DYS492=13, and, as the DYS492=13 were very few amongst the ancestor R-L11, this mutation happened amongst those few rather than the great majority with DYS492=12. It is also clear that from then the mutations forwards happen with a slower mutations rate because they start from 13 and not 12, or, better, the mutation from 13 to 14 happens with a faster mutation rate, but from 14 to 15 the mutation is much slower. The mutations forwards are faster than those downwards, but the values of 9 for R-P312 and 10 and 15 for R-U106 are next to the saturation.
Unfortunately we haven’t the data of SMGF about this marker. The mutation rate of DYS492 (Ballantyne) is very slow (3.92x10-4), then the lost of about 5% of the modal may also mean many thousands of years.

Funnily enough I agree with most of what you've written here, which doesn’t happen very often.

You can't say that U106 is younger than P312 from this data alone though, however I thing it probably is but not by much.

According to the relative STR variance calculations I've been doing the last couple of years, I always get U106 as younger than P312 as well.
Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #191 on: April 28, 2012, 11:21:43 PM »


Nothing is every ideal, which is what science tries to deal with. Everything I've read from Nordtvedt is that he is evaluating the data and trying to account for what he calls "fictions." That doesn't mean the mathematical models don't work.. it's a matter of precision.

Well tell me something I don’t already know. Ok, I don’t mean to be harsh, but yeah, the main problem here is the great number of assumptions that come into play on the “data” that Dr.Nordtvedt is evaluating. Like I said before, when we look at father-son’s pairs, we know what the allele at a given loci is, we know in that sense what the modal(ancestral) is, but even in that scenario there  could be cases of somatic mutations. As you can see in the study I provided, which used 23 autosomal STRs, which actually are a lot more stable(Which is self-evident in the case that even at the autosomal level, the paternal germ-line is three folds the maternal germline) than Y-chromosome STRs, there were 31 losses vs.22 gains. If one was to run a mathematical simulation across a generation you will never get a disparity of 9 mutations, at best 1-2. So again, this is family pairs, where all factors are known, and with a good precision we can determine mutation rates.

When it comes to the X Clan which has 100 participants, and a presumed common ancestor who lived 1300 ybp, the ancestral haplotype of that ancestor isn’t known, whether everyone in that generation is equally far removed, or at least on average far removed x generations from that ancestor isn’t know, the mutation rate is assumed to be constant, or to change very little, in order to estimate mutation rates in a time span of 1300 ybp. Often times the mutation rate will remain constant, because even with the fastest STRs the most mutations that will happen is 2 or 3 in 1300 ybp. However when one gets into 3000, 4000, 5000 ybp those fast mutating STRs probably incremented their mutation rate quote significantly along the process and one is presuming that 2*10^-3 measured for loci DYS-B from Clan X would do it, but in reality the average mutation rate could be something like 0.5*10^-3 for loci DYS-B when the time span is 3000-5000 ybp.  

I have no issue with your concern. I don't think your concern has a significant impact given the number of markers I'm evaluating. As I've shown on this thread, the results seem to show a correlation with the known SNP defined phylogenetic Y DNA tree, at least for the R1b subclade.  We know Ken thinks the correlation of STR diversity to time applies to the Hg I subclades appropriately as you can see from his web site.


... but I am NOT an expert so I'll request again:

Please, please, please make your challenges directly to Nordtvedt on the Rootsweb haplogroup I forum. I'm trying to be a proxy for some of your concerns, but you are much better at arguing your point of view than I am so please go ask Ken.

Of course you don’t think my concern will impact your work, of course none of the folks who are averaging mutation rates, and ignoring changes(loss of linearity) in mutation rate think it will have any impact in their work. If they did think so, then they wouldn’t do it, I don’t think anyone is going to waste a lot of precious time in something they know is wrong.  As for directing my challenges, I don’t want to challenge Dr.Nordtvedt, on the contrary, I prefer to run my own experiments on a case-control scenario, and see what results they yield.

This is not work, just a hobby so I'm not worrying at night too much about it.  I try to apply whatever models apply as best I can to the data I have - primarily P312 subclades, particularly L21.  That's it.  I've just found Dr. Nordtvedt's work impressive.

Contrary to your opinion, if you can come up with something specific I'll try to at least run some data through it and see what comes up. Based on Ironroads concerns and Busby's concerns about STR variance saturation, I took direction from Ironroads to understand what Heinila did to evaluate STR linearity. I implemented and use Heinila's work in the STR variance results I display. I've tried to normalize variance between STRs so no one STR outweighed another, but unfortunately I couldn't come up with a normalization technique that didn't throw out inconsistent results. Vince Vizachero has since explained that it was a difficult challenge and probably not worth the effort. I really don't care who's method does what, just whatever works best that I can implement (and of course, understand.)

If you don't want challenge Nordtvedt directly, that's fine, but why critique his work in absentia when you can do it directly with him?  If your arguments are strong, it will be apparent on the Rootsweb forum and we'll all know.

If you really want to progress the situation, please consider that if you convince him of the significance of your work then he will probably try to figure out how to improve existing models with your concept.  Everything I've seen from him is that he is sincere in making improvements to make a model that works. Wikipedia says he has shown a "a fiendish piece of mathematics" related to space work. I've seen the innovation he implemented with interclade calculations.  He's a guy you want on your side as he is adept at creating mathematical models.

« Last Edit: April 28, 2012, 11:33:06 PM by Mikewww » Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Maliclavelli
Guru
*****
Offline Offline

Posts: 2158


« Reply #192 on: April 29, 2012, 12:54:25 AM »

What have I to say to Ken Nordtvedt? I agree with you that he is a kind person. I have corresponded many times with him, firstly on Rootsweb and after my banishment on Dna-forums beyond many private letters. On Dna-forum I unmasked his nickname and he worried with some moderators. Finally I was banned also from there.

If we take the mutations of DYS492 we spoken above, we have:

3.92.10-4= 1 mutation every 2551 generations

The lost of the modal is about 5% in R-P312 and R-U106, calculating that there have been some back mutations.
(2551:100)x5=127.55

127.55x32=4081.6, i.e. about your dates.

What there is wrong in this?
What I have said about the outlier in R1a1a.
By doing this the outliers, which testify the mutations happened not overwhelmingly around the modal, are cut off, i.e. the haplotypes more “modal” or simply the haplotypes survived get the upper hand. If this isn’t important in a short period, it becomes important for a long one. See my letter above, where the aDNA found in Europe is underestimated for a 2.5 factor.

Here there is the mistake, that no mathematics may adjust.
« Last Edit: April 29, 2012, 12:55:55 AM by Maliclavelli » Logged

Maliclavelli


YDNA: R-S12460


MtDNA: K1a1b1e

JeanL
Old Hand
****
Offline Offline

Posts: 425


« Reply #193 on: April 29, 2012, 02:12:42 AM »

If you don't want challenge Nordtvedt directly, that's fine, but why critique his work in absentia when you can do it directly with him?  If your arguments are strong, it will be apparent on the Rootsweb forum and we'll all know.

When have I criticized Dr.Nordtvedt’s work directly? You have brought up certain assumptions made by folks on the hobbyist community, and I have explained my skepticism to some of them. I haven’t interacted much with Dr.Nordtvedt, but from what I have seen online, he seems like a reasonable, honest, humble person. I actually have argued before with Klyosov in other forums, and would gladly tell him straight to his face my concerns about his methodology, as I have done before. However, all I get from Klyosov is a cesspool of fallacious arguments, and the cowardice of posting fragments of our discussion in his Journal of Genetic Genealogy without giving the opportunity to defend my arguments, so I’m done with him, as to me, he is nothing but an arrogant …….
Logged
Jdean
Old Hand
****
Offline Offline

Posts: 678


« Reply #194 on: April 29, 2012, 07:23:15 AM »

I think that your comments are correct and timely.  As I said on my thread I will try to present data that shows that a large number of STR's are bounded in their range, which implies that hidden mutations are prevalent.  I think that hidden mutations will produce the same observable effect you cite, an apparent decrease in mutation rate over longer time spans.  I know the Maori and Gypsy data Zhiv used was not more than a thousand or so years, but if he was using the Faster mutators such as CDYa,b; then I would expect a similar observation would be made and apparently thats what he did?

I think it's at least reasonably clear that loci don't mutate equally up or down also it's likely that loci with different values will have slightly (probably very slightly) diffrent mutation rates.

However how big is this issue and do we need to worry about it that much, especially when talking about L11 and its offspring ?

DYS492 offers a convenient way to see if loci are likely to behave in a dramatically different way after a mutation.

looking at all the values for this loci in the P312 project I found

9         0.10%
10         0.21%
11         0.94%
12         95.60%
13         1.68%
14         1.47%

where as in the U106 project we see

10         0.12%
12         1.06%
13         95.89%
14         2.82%
15         0.12%


Both sets of data look remarkably similar apart from the obvious fact that 12 is modal in the first set and 13 in the second.

There definitely doesn’t seem to be any particular tendency for U106 to try and get back to a value of 12 and I think it's reasonable to conclude that if somebody who was P312 had a value of 13 at this loci that it would behave in exactly the same way as somebody who had 13 and was U106.

  The distribution about the modal is close in both sets of data.  Whether 5% of the mutations are multisteps is hard to discern.  Look at the behavior of 388 in R1b, I and J.  I think you'll be surprised?

Exactly but the modal is different in both cases, you seem to be suggesting (I think) that there is something special about the modal value itself that causes these distribution patterns but it's just normal distribution.

I will have a closer look at DYS388 but I don't think I will be terribly surprised, I just had a look at a small surname project that I knew had tight groups in J2b, U106, L21, I1 and E1b1b1. They had different values for that loci which isn't a great surprise considering the time to the common ancestor.


I've had a look at DYS388 in the I-L22 and P312 projects, it would have been pointless using the I1 project as that would have been comparing apples and oranges, L22 is thought to be about the same sort of age as P312 so should have similar variance.

P312

8      0.10%
10      0.10%
11      0.21%
12      97.41%
13      1.86%
14      0.31%

I-L22

12      0.60%
13      1.79%
14      93.45%
15      3.57%
16      0.60%


There are points of interest here but again both sets of data are remarkably similar, apart from the fact that the modal value for P312 is 12 and L22 14.
Logged

Y-DNA R-DF49*
MtDNA J1c2e
Kit No. 117897
Ysearch 3BMC9

Mark Jost
Old Hand
****
Offline Offline

Posts: 707


« Reply #195 on: April 29, 2012, 10:32:00 AM »

MikeW,

Have you ever tried to incorporate Chandlers mutation rate calculation formula into your spreadsheets to see there was differences in various Haplogroups or Subclades?

MJost

I don't apply mutation rates to any formulas in the Haplotype_Data spreadsheets i post. For any time calculation, I use whatever Ken Nordtvedt has in the Y DNA Tools Generations 7 methodology.

I asked if you ever tried as you have the sheer numbers.
Logged

148326
Pos: Z245 L459 L21 DF13**
Neg: DF23 L513 L96 L144 Z255 Z253 DF21 DF41 (Z254 P66 P314.2 M37 M222  L563 L526 L226 L195 L193 L192.1 L159.2 L130 DF63 DF5 DF49)
WTYNeg: L555 L371 (L9/L10 L370 L302/L319.1 L554 L564 L577 P69 L626 L627 L643 L679)
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #196 on: April 29, 2012, 11:23:54 AM »

What have I to say to Ken Nordtvedt? I agree with you that he is a kind person. I have corresponded many times with him, firstly on Rootsweb and after my banishment on Dna-forums beyond many private letters. On Dna-forum I unmasked his nickname and he worried with some moderators. Finally I was banned also from there.

If we take the mutations of DYS492 we spoken above, we have:

3.92.10-4= 1 mutation every 2551 generations

The lost of the modal is about 5% in R-P312 and R-U106, calculating that there have been some back mutations.
(2551:100)x5=127.55

127.55x32=4081.6, i.e. about your dates.

What there is wrong in this?
What I have said about the outlier in R1a1a.
By doing this the outliers, which testify the mutations happened not overwhelmingly around the modal, are cut off, i.e. the haplotypes more “modal” or simply the haplotypes survived get the upper hand. If this isn’t important in a short period, it becomes important for a long one. See my letter above, where the aDNA found in Europe is underestimated for a 2.5 factor.

Here there is the mistake, that no mathematics may adjust.

Why do you keep calculating TMRCA estimates based on one STR only?  Everything I read from the hobbyist scientists like Nordtvedt, Klyosov and Chandler, is that using more STRs (more experiments) is better. I think at the level of only one or two STRs, I'm not even sure it is worth doing.  The population genetics scientists even use at least ten or so, and they don't have much money so I'm sure they'd use more if they could.

Nordtvedt and Chandler are clear to say back-mutations are handled in their calcuations.

For R1a, you should probably take up this discussion with Anatole Klyosov. He is very interested in R1a, no doubt.  What TMRCA estimate (and by who) are you arguing against with your R1a1a example?

« Last Edit: April 29, 2012, 11:24:34 AM by Mikewww » Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #197 on: April 29, 2012, 11:30:15 AM »

I think that your comments are correct and timely.  As I said on my thread I will try to present data that shows that a large number of STR's are bounded in their range, which implies that hidden mutations are prevalent.  I think that hidden mutations will produce the same observable effect you cite, an apparent decrease in mutation rate over longer time spans.  I know the Maori and Gypsy data Zhiv used was not more than a thousand or so years, but if he was using the Faster mutators such as CDYa,b; then I would expect a similar observation would be made and apparently thats what he did?

I think it's at least reasonably clear that loci don't mutate equally up or down also it's likely that loci with different values will have slightly (probably very slightly) diffrent mutation rates.

However how big is this issue and do we need to worry about it that much, especially when talking about L11 and its offspring ?

DYS492 offers a convenient way to see if loci are likely to behave in a dramatically different way after a mutation.

looking at all the values for this loci in the P312 project I found

9         0.10%
10         0.21%
11         0.94%
12         95.60%
13         1.68%
14         1.47%

where as in the U106 project we see

10         0.12%
12         1.06%
13         95.89%
14         2.82%
15         0.12%


Both sets of data look remarkably similar apart from the obvious fact that 12 is modal in the first set and 13 in the second.

There definitely doesn’t seem to be any particular tendency for U106 to try and get back to a value of 12 and I think it's reasonable to conclude that if somebody who was P312 had a value of 13 at this loci that it would behave in exactly the same way as somebody who had 13 and was U106.

 The distribution about the modal is close in both sets of data.  Whether 5% of the mutations are multisteps is hard to discern.  Look at the behavior of 388 in R1b, I and J.  I think you'll be surprised?

Exactly but the modal is different in both cases, you seem to be suggesting (I think) that there is something special about the modal value itself that causes these distribution patterns but it's just normal distribution.

I will have a closer look at DYS388 but I don't think I will be terribly surprised, I just had a look at a small surname project that I knew had tight groups in J2b, U106, L21, I1 and E1b1b1. They had different values for that loci which isn't a great surprise considering the time to the common ancestor.


I've had a look at DYS388 in the I-L22 and P312 projects, it would have been pointless using the I1 project as that would have been comparing apples and oranges, L22 is thought to be about the same sort of age as P312 so should have similar variance.

P312

8      0.10%
10      0.10%
11      0.21%
12      97.41%
13      1.86%
14      0.31%

I-L22

12      0.60%
13      1.79%
14      93.45%
15      3.57%
16      0.60%


There are points of interest here but again both sets of data are remarkably similar, apart from the fact that the modal value for P312 is 12 and L22 14.


I understand what you are showing and I applaud the effort, but I still think an analysis of only one or two STRs is not enough to rely on.  There could be aberrations in any one STR in any one haplogroup. An example is U106's L1 subclade which has a lot of DYS439=null.

... but I think your point above for P312 and L22 is still valid.  Even though the modal for P312 and L22 are different, the dispersion is similar.    

This is just my perspective in looking at a glass of water and saying it is half full versus half empty, but what I think that people on this forum call "convergence around the modal" is really just divergence from the ancestral for scattered extant fairly young branches on the Y DNA tree.
« Last Edit: April 29, 2012, 07:00:29 PM by Mikewww » Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #198 on: April 29, 2012, 11:34:46 AM »

MikeW,

Have you ever tried to incorporate Chandlers mutation rate calculation formula into your spreadsheets to see there was differences in various Haplogroups or Subclades?

MJost

I don't apply mutation rates to any formulas in the Haplotype_Data spreadsheets i post. For any time calculation, I use whatever Ken Nordtvedt has in the Y DNA Tools Generations 7 methodology.

I asked if you ever tried as you have the sheer numbers.
I must not understand your question. Are you asking if I've run TMRCA calculations with multiple haplogroups? 

I've done that using Gen 7 for the major R-L11 haplogroups. That is displayed at the R-P312 Yahoo group Files under Haplogroup_Timeline_R-L11_Subclades.gif
Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Maliclavelli
Guru
*****
Offline Offline

Posts: 2158


« Reply #199 on: April 29, 2012, 11:52:56 AM »

Mikewww says: “Why do you keep calculating TMRCA estimates based on one STR only?  Everything I read from the hobbyist scientists like Nordtvedt, Klyosov and Chandler, is that using more STRs (more experiments) is better. I think at the level of only one or two STRs, I'm not even sure it is worth doing. The population genetics scientists even use at least ten or so, and they don't have much money so I'm sure they'd use more if they could.
Nordtvedt and Chandler are clear to say back-mutations are handled in their calcuations.

For R1a, you should probably take up this discussion with Anatole Klyosov. He is very interested in R1a, no doubt.  What TMRCA estimate (and by who) are you arguing against with your R1a1a example?”

Mine was of course only an experiment and I made many caveats. I did the same with the data posted about DYS388 and got different results, but I think that something interesting there is in these data.

My calculation about R1a1a has tried to demonstrate only that my results using the germ-line mutation rate weren’t different from those of the paper by using the Zhiv rate, this if we find an outlier like this. Of course the Zhiv rate was applied wrongly by those scholars, because they didn’t separate the haplotypes and the most diffused one falsified the calculation, otherwise with the Zhiv rate the age would have been at least 3 times older.
I have said here or in another thread that I am not against Nordtvedt, who said clearly that an hg I old of 6000 years had an interclade of 20000 years with another I. This demonstrates what we all know: that the most part of the lines are extinct.
« Last Edit: April 29, 2012, 11:56:36 AM by Maliclavelli » Logged

Maliclavelli


YDNA: R-S12460


MtDNA: K1a1b1e

Pages: 1 ... 6 7 [8] 9 10 ... 14 Go Up Print 
« previous next »
Jump to:  


SEO light theme by © Mustang forums. Powered by SMF 1.1.13 | SMF © 2006-2011, Simple Machines LLC

Page created in 0.116 seconds with 17 queries.