#130

As promised, I took a stab at aging the SNP blocks under Z16357. You can view the updated SNP chart with ages at? (you may need to hit Refresh to see the updates).

These are very rough estimates. The first date listed for each block is an estimate of the "formed date" - when the first SNP in that block was formed. The second date is a Time of Most Recent Common Ancestor (TMRCA) date - meaning when the last SNP in that block was formed, or the nearest date that two testers in child branches might expect to share a common ancestor in that block. We can't know which SNP in the block is oldest and which is newest or where within that date range common ancestors are actually shared (at least without additional documentation or testing).

Z16357 has several TMRCA estimates. YFull and SNP Tracker estimate it as?, Ytree (based on?Iain McDonald's work) has it as?, and Dave Vance (one of the L513 administrators) has it as . So that's 1000+ years difference between estimates. For broader range estimates, it's better to assume an older age.

Our 28 known testers have an average of 55 variants at or below Z16357 - a high of 67 and low of 45 (st. dev. of 5.7). Much of this variability is due to higher variant count coverage/possibilities for Big-Y 700 testers vs. Big-Y 500 testers - I didn't provide any factoring/weighting for the different test types (I'm not certain who has which test type anyway).

I also did not consider STR differences - which could provide some insight into larger blocks. For example, if an SNP block spans hundreds of years, STR differences could give insight into whether the MRCA is at the beginning or end of that time span.?I did, however, make some minor adjustments in time spans for the few known common ancestors we have across testers.

With a formed age estimate of 3900 years ago, this comes out to 71 years per SNP - well faster than the generally-accepted 83.3 value used on many aging calculations. I'm using FTDNA's average aggregated private variants for each terminal block (rather than exact private variant values). With this 71 years value, I then stepped backward from 1950AD to 1950BC (the 3900 year span) to assign date estimates for each block.

You can review my spreadsheet for the calculations at The spreadsheet also provides TMRCA estimates if Z16357 instead has formed dates of 3500 years ago (63.7 years per SNP) or 3000 years ago (54.6 years per SNP). These differences are relatively minor for more recent haplogroups, but become quite significant for older haplogroups.

One item of note is the significant time distance between the two sides of the Z16357 tree - Z16343 and BY19970 split probably 3000+ years ago and then each had 2000+ years with no (yet) known branches - which is primarily why most of my research analyzes these in isolation from each other.

I'd be VERY happy to have feedback or criticism on my methodologies or the dates estimated. I'll update these estimates any time we have a notable update to the tree.

Cheers,

Jared

#131

开云体育

Hi Jared,

?

Thanks for your thorough work. It may be interesting for you to share your findings on the R1b-L513 Facebook page to see what some of the other experts think there also.

?

I have not looked at your methodology in detail, but in general the concept seems very good. Tiger Walsh has stated on, I believe, the L513 Facebook page that older SNPs can be in the range of 65 to 75 years per SNP. That is because when there are thousands of testers for those older SNPs, more irregular SNPs or SNPs in poorer regions can be verified. Although we don’t have thousands of testers, perhaps we are starting to see this in our small group. Or we just are producing more than the average SNPs. Which is good for branching and dating.

?

I have looked mostly at just my small branch of Hartleys. In that branch I would say that 1650 is in my opinion late for a common ancestor date of A11132. I see how you got that date. My branch with my brother has 7 SNPs up to A11132 and the branch with Steve and Michael Hartley have an average of only two SNPs. However, their actual individual Private Variants are 5 and 2. I am told that is because some of their individual Private Variants are not from trusted areas of tested YDNA. My guess for the date of the common ancestor of A11132 Hartleys is 1500 based on looking at the STRs and SNPs. However, I am looking forward to the BigY testing of two more Hartleys to further refine dates and branching. Your date of 900 for the formation of A11132 is interesting also as it implies that there should be other non-Hartleys in this group that we have not found yet.

?

Joel

?

Show quoted text

From: [email protected] On Behalf Of Jared Smith
Sent: Monday, January 3, 2022 8:45 PM
To: [email protected]
Subject: [Z16357] TMRCA aging of Z16357 haplogroups

?

As promised, I took a stab at aging the SNP blocks under Z16357. You can view the updated SNP chart with ages at? (you may need to hit Refresh to see the updates).

These are very rough estimates. The first date listed for each block is an estimate of the "formed date" - when the first SNP in that block was formed. The second date is a Time of Most Recent Common Ancestor (TMRCA) date - meaning when the last SNP in that block was formed, or the nearest date that two testers in child branches might expect to share a common ancestor in that block. We can't know which SNP in the block is oldest and which is newest or where within that date range common ancestors are actually shared (at least without additional documentation or testing).

Z16357 has several TMRCA estimates. YFull and SNP Tracker estimate it as?, Ytree (based on?Iain McDonald's work) has it as?, and Dave Vance (one of the L513 administrators) has it as . So that's 1000+ years difference between estimates. For broader range estimates, it's better to assume an older age.

Our 28 known testers have an average of 55 variants at or below Z16357 - a high of 67 and low of 45 (st. dev. of 5.7). Much of this variability is due to higher variant count coverage/possibilities for Big-Y 700 testers vs. Big-Y 500 testers - I didn't provide any factoring/weighting for the different test types (I'm not certain who has which test type anyway).

?

I also did not consider STR differences - which could provide some insight into larger blocks. For example, if an SNP block spans hundreds of years, STR differences could give insight into whether the MRCA is at the beginning or end of that time span.?I did, however, make some minor adjustments in time spans for the few known common ancestors we have across testers.

?

With a formed age estimate of 3900 years ago, this comes out to 71 years per SNP - well faster than the generally-accepted 83.3 value used on many aging calculations. I'm using FTDNA's average aggregated private variants for each terminal block (rather than exact private variant values). With this 71 years value, I then stepped backward from 1950AD to 1950BC (the 3900 year span) to assign date estimates for each block.

?

You can review my spreadsheet for the calculations at The spreadsheet also provides TMRCA estimates if Z16357 instead has formed dates of 3500 years ago (63.7 years per SNP) or 3000 years ago (54.6 years per SNP). These differences are relatively minor for more recent haplogroups, but become quite significant for older haplogroups.

?

One item of note is the significant time distance between the two sides of the Z16357 tree - Z16343 and BY19970 split probably 3000+ years ago and then each had 2000+ years with no (yet) known branches - which is primarily why most of my research analyzes these in isolation from each other.

?

I'd be VERY happy to have feedback or criticism on my methodologies or the dates estimated. I'll update these estimates any time we have a notable update to the tree.

?

Cheers,

?

Jared

?

#132

Great stuff! Thank you.?

Show quoted text

On Mon, Jan 3, 2022 at 8:45 PM Jared Smith <jared@...> wrote:

As promised, I took a stab at aging the SNP blocks under Z16357. You can view the updated SNP chart with ages at? (you may need to hit Refresh to see the updates).

These are very rough estimates. The first date listed for each block is an estimate of the "formed date" - when the first SNP in that block was formed. The second date is a Time of Most Recent Common Ancestor (TMRCA) date - meaning when the last SNP in that block was formed, or the nearest date that two testers in child branches might expect to share a common ancestor in that block. We can't know which SNP in the block is oldest and which is newest or where within that date range common ancestors are actually shared (at least without additional documentation or testing).

Z16357 has several TMRCA estimates. YFull and SNP Tracker estimate it as?, Ytree (based on?Iain McDonald's work) has it as?, and Dave Vance (one of the L513 administrators) has it as . So that's 1000+ years difference between estimates. For broader range estimates, it's better to assume an older age.

Our 28 known testers have an average of 55 variants at or below Z16357 - a high of 67 and low of 45 (st. dev. of 5.7). Much of this variability is due to higher variant count coverage/possibilities for Big-Y 700 testers vs. Big-Y 500 testers - I didn't provide any factoring/weighting for the different test types (I'm not certain who has which test type anyway).

I also did not consider STR differences - which could provide some insight into larger blocks. For example, if an SNP block spans hundreds of years, STR differences could give insight into whether the MRCA is at the beginning or end of that time span.?I did, however, make some minor adjustments in time spans for the few known common ancestors we have across testers.

With a formed age estimate of 3900 years ago, this comes out to 71 years per SNP - well faster than the generally-accepted 83.3 value used on many aging calculations. I'm using FTDNA's average aggregated private variants for each terminal block (rather than exact private variant values). With this 71 years value, I then stepped backward from 1950AD to 1950BC (the 3900 year span) to assign date estimates for each block.

You can review my spreadsheet for the calculations at The spreadsheet also provides TMRCA estimates if Z16357 instead has formed dates of 3500 years ago (63.7 years per SNP) or 3000 years ago (54.6 years per SNP). These differences are relatively minor for more recent haplogroups, but become quite significant for older haplogroups.

One item of note is the significant time distance between the two sides of the Z16357 tree - Z16343 and BY19970 split probably 3000+ years ago and then each had 2000+ years with no (yet) known branches - which is primarily why most of my research analyzes these in isolation from each other.

I'd be VERY happy to have feedback or criticism on my methodologies or the dates estimated. I'll update these estimates any time we have a notable update to the tree.

Cheers,

Jared

#133

Thank you so much for the awesome information.

Show quoted text

-----Original Message-----
From: Jared Smith <jared@...>
To: [email protected]
Sent: Mon, Jan 3, 2022 8:44 pm
Subject: [Z16357] TMRCA aging of Z16357 haplogroups

As promised, I took a stab at aging the SNP blocks under Z16357. You can view the updated SNP chart with ages at? (you may need to hit Refresh to see the updates).

These are very rough estimates. The first date listed for each block is an estimate of the "formed date" - when the first SNP in that block was formed. The second date is a Time of Most Recent Common Ancestor (TMRCA) date - meaning when the last SNP in that block was formed, or the nearest date that two testers in child branches might expect to share a common ancestor in that block. We can't know which SNP in the block is oldest and which is newest or where within that date range common ancestors are actually shared (at least without additional documentation or testing).

Z16357 has several TMRCA estimates. YFull and SNP Tracker estimate it as?, Ytree (based on?Iain McDonald's work) has it as?, and Dave Vance (one of the L513 administrators) has it as . So that's 1000+ years difference between estimates. For broader range estimates, it's better to assume an older age.

Our 28 known testers have an average of 55 variants at or below Z16357 - a high of 67 and low of 45 (st. dev. of 5.7). Much of this variability is due to higher variant count coverage/possibilities for Big-Y 700 testers vs. Big-Y 500 testers - I didn't provide any factoring/weighting for the different test types (I'm not certain who has which test type anyway).

I also did not consider STR differences - which could provide some insight into larger blocks. For example, if an SNP block spans hundreds of years, STR differences could give insight into whether the MRCA is at the beginning or end of that time span.?I did, however, make some minor adjustments in time spans for the few known common ancestors we have across testers.

With a formed age estimate of 3900 years ago, this comes out to 71 years per SNP - well faster than the generally-accepted 83.3 value used on many aging calculations. I'm using FTDNA's average aggregated private variants for each terminal block (rather than exact private variant values). With this 71 years value, I then stepped backward from 1950AD to 1950BC (the 3900 year span) to assign date estimates for each block.

You can review my spreadsheet for the calculations at The spreadsheet also provides TMRCA estimates if Z16357 instead has formed dates of 3500 years ago (63.7 years per SNP) or 3000 years ago (54.6 years per SNP). These differences are relatively minor for more recent haplogroups, but become quite significant for older haplogroups.

One item of note is the significant time distance between the two sides of the Z16357 tree - Z16343 and BY19970 split probably 3000+ years ago and then each had 2000+ years with no (yet) known branches - which is primarily why most of my research analyzes these in isolation from each other.

I'd be VERY happy to have feedback or criticism on my methodologies or the dates estimated. I'll update these estimates any time we have a notable update to the tree.

Cheers,

Jared

#134

Thanks for your feedback and thoughts Joel! A few comments are below.

It may be interesting for you to share your findings on the R1b-L513 Facebook page to see what some of the other experts think there also.?

Yes, I'll share soon. It sounds like some things may change in the near future as we get some additional test results in.

?

Tiger Walsh has stated on, I believe, the L513 Facebook page that older SNPs can be in the range of 65 to 75 years per SNP.

Yes, this aggregating?effect for SNP counts based?on multiple testers increasing?reliability of reads is known,?but is a bit tricky to account for. We're already well in or even faster than that 65 to 75 range with both old and new SNPs - and across both Y700 and Y500 testers. My hope is to eventually apply factoring for "years per SNP" based on age of the SNP and whether it's a 700 or 500 test to account for much of this. I'd also like to calculate confidence intervals on the various age estimates. James Irvine has an excellent, recently-published journal article that covers these TMRCA topics in depth -?
?

Or we just are producing more than the average SNPs. Which is good for branching and dating.

Everything seems to suggest that our branches have notably higher SNP counts than most. This is very useful for us for near-term age estimates and branch generation, but it does make aging more tricky when our SNP counts don't align with the predicted "formed" dates for our old haplogroups - it's difficult to know whether we're actually making more SNPs or whether Z16357 is just much older than predicted.

?

However, their actual individual Private Variants are 5 and 2. I am told that is because some of their individual Private Variants are not from trusted areas of tested YDNA.

Yes, this is essentially the opposite problem of the effect discussed above.?Alex's Big Tree looks more closely at these private variants. 4 years ago I did the same with raw results for most of the folks then in our group, but with a lot more people now and the transition since then to hg38 for raw data, I'm not sure if I want to take on this rather complex kit-by-kit analysis again. Doing so would require everyone to provide me their?updated raw BigY data (or administering an FTDNA project in which everyone is a member). For now, I'm prone to use FTDNAs conservative estimates for private variants. With the opposite effect for older haplogroups, perhaps this is a bit of a wash overall?

Supposedly FTDNA is going to provide their own age estimates one of these years. We'll see.

?

Your date of 900 for the formation of A11132 is interesting also as it implies that there should be other non-Hartleys in this group that we have not found yet.

It should also imply that there would be others more closely related to me - yet, alas, I'm still stuck rather isolated at the ~1000 year old A11138. These types of long branches that were seemingly isolated for many generations are really the story of much of Z16357 - it's what makes us fairly unique and what makes me so interested in solving some of these puzzles.

Cheers,

Jared