FTDNA's New Time Tree

#131

FTDNA has added a "Time Tree" feature to it's Discover More tool.? It reminds me of the Scientific View in YFull, You may enjoy checking it out.

-Tom LIttle
Little Project

#132

I was remiss by not adding the FTDNA's new paper on aging:

#133

Interesting short summary on the announcement but it is sparse on the underlying details of the algorithms themselves... The 'relaxed' method simply sounds like a means to 'force-fit' apparently confusing SNP stacks of one family lineage against another brother line under the same genetic cluster (branch node).? The whole discussion appears to contradict Iain McDonald's contention that a "strict" clock applies to all.

leake

#134

I haven't done an indepth study of the above TMRCA methods, and this is the first of heard of a 'strict clock' which I doubt is possible. I have checked Ian's TMRCAs, and didn't find strictness, if I'm understanding it's meaning. I have a few questions for him, like "if you're using 160 yrs/SNP then why do I sometimes only find ~60 years in between SNPs that are accepted as only having one SNP between the father SNP 'block' and son SNP (no equivalents)"?

Concerning brother SNPs and force fitting, I don't think it works well in some haplogroups like M222 which can sometimes have extremely more equivalents than their brother subclades. Yfull may be on the right track here where they just average all the brothers, and keep doing so as far back in time as possible. L513 may be lucky in that Yfull's # of SNPs 'average' for all son SNPs of DF13 seemed to contain the exact average number of SNPs within it.?

Here's something I've been pondering lately: the 'daughtering out' of male lineages, which I think, but not sure, is 9/10 male lineages eventually disappear due to no males being born, which is called daughtering out. I wonder if the gifted mathematicians have included "daughtering out" in their algorithms.? In the current time period I suspect some of our male lineages, like mine, may have a skinny appearance to them, especially going back in time for the last 700 years or so. Skinny would mean not enough son's having son issue. My own g grandfather was the single male, + one sister, born to my gg grandfather who died before his 21st birthday.

When I see "bushy" lineages, like some in M222 with many more SNPs than their brother SNPs, I think 'rich' vs 'not so rich' and/or 'poor'. If everything is passed down to the oldest son this makes it more difficult for the other sons to thrive. While not true in all cases, the richest can afford to have more issue, and if there are big titles attached to their names they will want plenty of sons in order to increase their riches and influence for hundreds of years to come. It's likely not as simple as just riches, but gamblers looking for an 'edge' will tell you that any edge above 50% is very exciting for them to exploit. So if the prince is handed a kingdom his chances of successfully exploiting it to his lineage's advantage is greater than his younger brothers. What's the algorithm for that? ;-)

Daryl?

Show quoted text

On Sat, Sep 10, 2022, 7:28 AM Leake Little <leakelittle@...> wrote:

Interesting short summary on the announcement but it is sparse on the underlying details of the algorithms themselves... The 'relaxed' method simply sounds like a means to 'force-fit' apparently confusing SNP stacks of one family lineage against another brother line under the same genetic cluster (branch node).? The whole discussion appears to contradict Iain McDonald's contention that a "strict" clock applies to all.

leake

#135
Edited

The default way to do TMRCA estimates is use predetermined mutation rate. You can all this a strict clock or whatever? you want, but as far as we know (the scientific studies to date), the Y chromosome mutates basically at the same rate across geographies, ethnicities, subclades.

We know from reality, that some subclades are different from other. That doesn't change the expected rate of mutation, though.?? You can flip a coin five times and get five heads in a row, but the expected rate of heads/tales for the next flip is still 50/50. Still five heads in a row can come up.

Therefore, if you have a subclade with a known difference, there are mathematical models that allow you introduce a constraint. You can call this a force fit if you want, but it basically says a certain SNP was around at a certain regardless of what the expected rate of mutations are. What FTDNA is calling a relaxed clock is just this. It is the introducing of reality points into a statisical model.

When I first learned computer software development, I had to write linear programming software and computer simulations. There are various math models that all of these things are based on. FTDNA is just using them, with the guidance of Dr. Iain McDonald. To me, it is obvious logical thing for FTDNA to do.

Unfortunately, YFull does not have an advanced math model so rather than use linear programming techniques they essentially just fudge their TMRCAs when the expected rates don't produce results that account for reality (like ancient DNA finds).

I can't discuss this fully on multiple forums but you can talk to Dr. McDonald on the U106 groups.io.? I'm tracking updates and sharing what I find on Anthrogenica under FTDNA.

I've said this for a number of years as I use to maintain STR based TMRCA spreadsheets.
----?? TMRCAs are no panacea. If a method produces a result that fits your hypothesis or meets what you are otherwise are accustomed to,? you are likely to think that TMRCA method is the greatest things since sliced bread. If it doesn't fit, then you will think it is worthless.? ----

One thing we all can do is update our Family Trees on our kit dashboards on FTDNA and link our kits and known cousin kits to people on the family tree. Make sure to include birth dates. FTDNA is using this data for BIg Y testers as validation for their TMRCA methodology.? This is still all in beta, so more changes will be made.? FTDNA will eventually expand their FTBED regional defintions, which math-wise, should provide more accuracy.

#136

Daryl,

You need to git outta my noggin '!

Well stated sir?

Brad?

Show quoted text

On Sat, Sep 10, 2022, 14:42 Class1 Driver <class1driver@...> wrote:

I haven't done an indepth study of the above TMRCA methods, and this is the first of heard of a 'strict clock' which I doubt is possible. I have checked Ian's TMRCAs, and didn't find strictness, if I'm understanding it's meaning. I have a few questions for him, like "if you're using 160 yrs/SNP then why do I sometimes only find ~60 years in between SNPs that are accepted as only having one SNP between the father SNP 'block' and son SNP (no equivalents)"?

Concerning brother SNPs and force fitting, I don't think it works well in some haplogroups like M222 which can sometimes have extremely more equivalents than their brother subclades. Yfull may be on the right track here where they just average all the brothers, and keep doing so as far back in time as possible. L513 may be lucky in that Yfull's # of SNPs 'average' for all son SNPs of DF13 seemed to contain the exact average number of SNPs within it.?

Here's something I've been pondering lately: the 'daughtering out' of male lineages, which I think, but not sure, is 9/10 male lineages eventually disappear due to no males being born, which is called daughtering out. I wonder if the gifted mathematicians have included "daughtering out" in their algorithms.? In the current time period I suspect some of our male lineages, like mine, may have a skinny appearance to them, especially going back in time for the last 700 years or so. Skinny would mean not enough son's having son issue. My own g grandfather was the single male, + one sister, born to my gg grandfather who died before his 21st birthday.

When I see "bushy" lineages, like some in M222 with many more SNPs than their brother SNPs, I think 'rich' vs 'not so rich' and/or 'poor'. If everything is passed down to the oldest son this makes it more difficult for the other sons to thrive. While not true in all cases, the richest can afford to have more issue, and if there are big titles attached to their names they will want plenty of sons in order to increase their riches and influence for hundreds of years to come. It's likely not as simple as just riches, but gamblers looking for an 'edge' will tell you that any edge above 50% is very exciting for them to exploit. So if the prince is handed a kingdom his chances of successfully exploiting it to his lineage's advantage is greater than his younger brothers. What's the algorithm for that? ;-)

Daryl?

On Sat, Sep 10, 2022, 7:28 AM Leake Little <leakelittle@...> wrote:
Interesting short summary on the announcement but it is sparse on the underlying details of the algorithms themselves... The 'relaxed' method simply sounds like a means to 'force-fit' apparently confusing SNP stacks of one family lineage against another brother line under the same genetic cluster (branch node).? The whole discussion appears to contradict Iain McDonald's contention that a "strict" clock applies to all.

leake

#137

Thanks for posting these, Tom.

On Saturday, September 10, 2022 at 09:28:54 AM CDT, Leake Little <leakelittle@...> wrote:

Interesting short summary on the announcement but it is sparse on the underlying details of the algorithms themselves... The 'relaxed' method simply sounds like a means to 'force-fit' apparently confusing SNP stacks of one family lineage against another brother line under the same genetic cluster (branch node).? The whole discussion appears to contradict Iain McDonald's contention that a "strict" clock applies to all.

leake

#138

That's true, Daryl, "daughtering out" does have a profound effect on a lineage. Something I've noticed, while searching for genealogy threads, is that many rich land owners would also include their illegitimate children in their wills, which I initially found to be a bit awkward. I wondered if by acknowledging these children, there was a better chance that their lineage would continue, even through a bastard son. Royalty obviously did it, but I would say lesser land owners could do it as well, and for the same reasons. I have a particular affinity for wills because of the wealth of information they can provide. Unfortunately, these can be a rare find, especially the further back in time you go.

pk

On Saturday, September 10, 2022 at 04:42:47 PM CDT, Class1 Driver <class1driver@...> wrote:

I haven't done an indepth study of the above TMRCA methods, and this is the first of heard of a 'strict clock' which I doubt is possible. I have checked Ian's TMRCAs, and didn't find strictness, if I'm understanding it's meaning. I have a few questions for him, like "if you're using 160 yrs/SNP then why do I sometimes only find ~60 years in between SNPs that are accepted as only having one SNP between the father SNP 'block' and son SNP (no equivalents)"?

Concerning brother SNPs and force fitting, I don't think it works well in some haplogroups like M222 which can sometimes have extremely more equivalents than their brother subclades. Yfull may be on the right track here where they just average all the brothers, and keep doing so as far back in time as possible. L513 may be lucky in that Yfull's # of SNPs 'average' for all son SNPs of DF13 seemed to contain the exact average number of SNPs within it.?

Here's something I've been pondering lately: the 'daughtering out' of male lineages, which I think, but not sure, is 9/10 male lineages eventually disappear due to no males being born, which is called daughtering out. I wonder if the gifted mathematicians have included "daughtering out" in their algorithms.? In the current time period I suspect some of our male lineages, like mine, may have a skinny appearance to them, especially going back in time for the last 700 years or so. Skinny would mean not enough son's having son issue. My own g grandfather was the single male, + one sister, born to my gg grandfather who died before his 21st birthday.

When I see "bushy" lineages, like some in M222 with many more SNPs than their brother SNPs, I think 'rich' vs 'not so rich' and/or 'poor'. If everything is passed down to the oldest son this makes it more difficult for the other sons to thrive. While not true in all cases, the richest can afford to have more issue, and if there are big titles attached to their names they will want plenty of sons in order to increase their riches and influence for hundreds of years to come. It's likely not as simple as just riches, but gamblers looking for an 'edge' will tell you that any edge above 50% is very exciting for them to exploit. So if the prince is handed a kingdom his chances of successfully exploiting it to his lineage's advantage is greater than his younger brothers. What's the algorithm for that? ;-)

Daryl?

Show quoted text

On Sat, Sep 10, 2022, 7:28 AM Leake Little <leakelittle@...> wrote:

Interesting short summary on the announcement but it is sparse on the underlying details of the algorithms themselves... The 'relaxed' method simply sounds like a means to 'force-fit' apparently confusing SNP stacks of one family lineage against another brother line under the same genetic cluster (branch node).? The whole discussion appears to contradict Iain McDonald's contention that a "strict" clock applies to all.

leake

开云体育

FTDNA's New Time Tree