Mutations : Don't mistake real rates with high-qual rates

Any discussions regarding y-DNA markers, results or questions.

Posts: 143
Joined: Sun Mar 18, 2012 6:26 pm
Location: Paris region
YDNA:
G2a2b2a1b1a2a-CT4803
MtDNA:
H2a2a1 (rCRS)
PostPosted: Fri Sep 26, 2014 3:13 pm
On April,29th 2014, I wrote a post in Anthrogenica

http://www.anthrogenica.com/showthread. ... #post38720

Here are the contents
"Yesterday,in https://www.facebook.com/groups/9995363812/ there is a very interesting comparizon published by Itaï Juste:
"I just compared FG and BIG Y results from Mr. Kraus in my project who did both tests, using FG interpretation for both results, so they can be compared on the same bases.
As a reminder, here is FG interpretation reliability scale:
no flag: over 99% likely genuine (95% for INDELs); *: over 95% likely genuine (90% for INDELs); **: about 40% likely genuine; ***: about 10% likely genuine

I get following statistics:
SNP in KRAUS FG: 5504; the no flag: 905; *: 254; **: 1710; ***: 2635

SNP in KRAUS BIG Y: 4925;_ no flag: 554; *: 125; **: 858; ***: 3388

SNP in both KRAUS FG and BIG Y: 1879.

no flag in FG : no flag in BIG Y: 486; * in BIG Y: 80; ** in BIG Y: 51; *** in BIG Y: 33.

* in FG :_______no flag in BIG Y: 49; * in BIG Y: 31; ** in BIG Y: 12; *** in BIG Y: 13.

** in FG;______no flag in BIG Y: 15; * in BIG Y: 12; ** in BIG Y: 412; *** in BIG Y: 297.

*** in FG______no flag in BIG Y: 2;_ * in BIG Y: 1;_ ** in BIG Y: 106; *** in BIG Y: 279

Different called results in BIG Y and FG (one positive and the other negative): 374+9
no flag in BIG Y: 0; * in BIG Y: 0; ** in BIG Y: 43; *** in BIG Y: 340
"

I think about 90-99% of SNPs are found by the FGC analysis for FGC/BGI raw data (about 15 Mbp selected) and Big Y data (about 10 Mbp selected) but hardly the half of the real found markers are with high qualifications.

For about 10M analysed in BIG Y raw data, there are 1879 found in FGC/BGI raw data, almost every are correct and the real number of SNPs should be about 1900 for 10M and 2850 for 15M.

The proportion of high qualifications for FGC data is about 1159/2850=40,66% and for BIG Y data about 679/1900=35,75%
If we suppose the reference is the Adam of the split of A0 and A1 (nowithstanding of A00) which dates from 160-240,000 years

This gives for
- FGC data a mutation from every 160,000/1159 = 138 years to every 240,000/1159 = 207 years.
- Big Y data a mutattion from every 160,000/679 = 235 years to every 240,000/679 = 353 years.

My preference is for the high datation which allows an "Out of Africa" during the very favourable, very warm and very wet Eemian (130-125,000 years ago) and I will take for estimation for FGC data 1 high-qualification mutation every 200 years (5 mutations every millenium) and for BIG Y data 1 high-qualification mutation every 333 years (3 mutations every millenium).

40% gives a real rate of about 10/4= 2.5 mutations every 200 years or 1 real mutation every 200/2.5= 80years for 15M or 1 mutation every 120 years for 10M."

My goal was to well marked the number differences between real mutations and High Quality mutations because I found a lot of anthrogenica participants didn't do the difference and this involves a trouble in the calculation of mutation rates. In more, I found this absence of discrimination was not innocent and allowed to make easy the validation of their theories about R1b=Bell Beakers and to avoid a big difference with the dates calculated from the STR diversity which were their old reference in their theories.
I found
1) the High Qualty mutations were about 40% of the real mutations for FGC (in 15M bp) and about 36% for BigY (in 10Mb)

2)
a) If Adam lived 160,000 years ago, for FGC (15Mbp) , the high-quality mutation rate was 1 mutation every 138 years , the real mutation rate was 138x40,66/100= 56 years for 15Mb.
If Adam lived 160,000 years ago, for BIG Y (10Mbp), the high-quality mutation rate was 1 mutation every 235 years , the mutation rate was 235x35,75/100= 84 years. for 10Mb

b) If Adam lived 2400,000 years ago, for FGC (15Mbp) , the high-quality mutation rate was 1 mutation every 207 years , the real mutation rate was 207x40,66/100= 84 years for 15Mb.
If Adam lived 160,000 years ago, for BIG Y (10Mbp), the high-quality mutation rate was 1 mutation every 353 years , the mutation rate was 353x35,75/100= 126 years. for 10Mb


Therefore I found 1 real mutation rate every 56-84 years for FGC (15M) and every 84-126 years for Big Y (10M), and not the same thing, 1 high qual. mutation every 138-207 years for FGC (15M) and every 235-353 years for Big Y (10M)

Posts: 623
Joined: Wed Mar 14, 2012 2:16 pm
PostPosted: Sun Sep 28, 2014 3:17 am
palamede wrote:If we suppose the reference is the Adam of the split of A0 and A1 (nowithstanding of A00) which dates from 160-240,000 years

This gives for
- FGC data a mutation from every 160,000/1159 = 138 years to every 240,000/1159 = 207 years.
- Big Y data a mutattion from every 160,000/679 = 235 years to every 240,000/679 = 353 years.

My preference is for the high datation

You have chosen a mutation rate to fit your own hypothesis.

Posts: 2410
Joined: Sun Mar 18, 2012 7:08 am
Location: Pisa (Italy)
YDNA:
R- Z2110 (KV7Y2)
MtDNA:
K1a1b1e/HQ176413
PostPosted: Sun Sep 28, 2014 5:23 am
lgmayka wrote:You have chosen a mutation rate to fit your own hypothesis.


To use some hypotheses about the mutation rate to fit our own theories is part of any scientific method, which does hypotheses based upon some inferences and after tries the experiments to demonstrate that. But what is resolutive in the science are the proofs.
I said that probably amongst the "Italian" agriculturalists who migrated by sea to Iberia 7500 years ago there were already some R-L51, because we find this haplogroup now in the place where they landed beyond in Italy, and specifically Tuscany (the basal L51 is in one Tuscan 1KGP sample so far). We have also a link of this ancestral subclade of L51 in Tuscany and in Ireland, other place where those migrants landed from Iberia, and another more recent subclade in Central Europe till Britannia.
Of course to all this we have to find proofs in the aDNA.

Anyway the mutation rate of the STRs you all used is now definitely a flaw, and about this no further proof we need.

Posts: 143
Joined: Sun Mar 18, 2012 6:26 pm
Location: Paris region
YDNA:
G2a2b2a1b1a2a-CT4803
MtDNA:
H2a2a1 (rCRS)
PostPosted: Sun Sep 28, 2014 9:44 am
lgmayka wrote:
palamede wrote:If we suppose the reference is the Adam of the split of A0 and A1 (nowithstanding of A00) which dates from 160-240,000 years

This gives for
- FGC data a mutation from every 160,000/1159 = 138 years to every 240,000/1159 = 207 years.
- Big Y data a mutattion from every 160,000/679 = 235 years to every 240,000/679 = 353 years.

My preference is for the high datation

You have chosen a mutation rate to fit your own hypothesis.


It is not a real critic, It's too easy to accuse of wishfull thinking without examining the logic (right or bad)of the arguments.

More than criticizing my preference based on the hypotheses accepted by a part of the scientific world of an Out of Africa dating of the Eemian (130,000-115,000BP), I wished my reasoning was challenged.
Is my comparaison of the results of the 2 test FGC and BIG Y of one user and my conclusion about the mutation rates and the ratio differences between real mutations and HQ mutations are right or not ?

Anyway, in my conclusion of September, I avoided to take my preference again and I preferred to give a range with Adam date from 160,000 years to 240,000 years ago.

It seems me the necessary discussion about the ratio is bypassed, I dont know why and why there is a strong trend to say real mutations = HQ mutation . This could not be true and I don't understand why there is no discussion about the ratio while there are data to determine it and why my tentative trial is ignored.

but if you don't follow the apparent consensus, you are despised and ignored, nobody try to refute your arguments. But there were a lot of in fashion theories which have disappeared and they would be disappeared quicker, if their noisy bearers had examined the arguments of their opponents carefully .

Posts: 34
Joined: Thu Mar 15, 2012 3:54 am
PostPosted: Mon Sep 29, 2014 11:07 am
palamede wrote:
lgmayka wrote:
palamede wrote:If we suppose the reference is the Adam of the split of A0 and A1 (nowithstanding of A00) which dates from 160-240,000 years

This gives for
- FGC data a mutation from every 160,000/1159 = 138 years to every 240,000/1159 = 207 years.
- Big Y data a mutattion from every 160,000/679 = 235 years to every 240,000/679 = 353 years.

My preference is for the high datation

You have chosen a mutation rate to fit your own hypothesis.


It is not a real critic, It's too easy to accuse of wishfull thinking without examining the logic (right or bad)of the arguments.

More than criticizing my preference based on the hypotheses accepted by a part of the scientific world of an Out of Africa dating of the Eemian (130,000-115,000BP), I wished my reasoning was challenged.
Is my comparaison of the results of the 2 test FGC and BIG Y of one user and my conclusion about the mutation rates and the ratio differences between real mutations and HQ mutations are right or not ?

Anyway, in my conclusion of September, I avoided to take my preference again and I preferred to give a range with Adam date from 160,000 years to 240,000 years ago.

It seems me the necessary discussion about the ratio is bypassed, I dont know why and why there is a strong trend to say real mutations = HQ mutation . This could not be true and I don't understand why there is no discussion about the ratio while there are data to determine it and why my tentative trial is ignored.

but if you don't follow the apparent consensus, you are despised and ignored, nobody try to refute your arguments. But there were a lot of in fashion theories which have disappeared and they would be disappeared quicker, if their noisy bearers had examined the arguments of their opponents carefully .

Did you take into account HQ area defined by BED file?

Return to y-DNA Discussions

Who is online

Users browsing this forum: Google [Bot] and 0 guests