PDA

View Full Version : Hypothesis: MyOrigins percentages biased towards y DNA haplogroup



sktibo
11-14-2016, 12:09 AM
Hello, this is something of an experiment,

I believe that our Y haplogroup greatly influences our MyOrigins percentages, and I'm hoping more people will chime in to help prove or disprove my idea.

To start, in my own case, I am about 65% British isles 35% mainland Europe ancestry on paper, but I got 36% B/I, 51% west and central Europe, (11% Scandinavia also). My most distant traceable paternal ancestor was from West Central France.

Edit: turns out my hypothesis was incorrect. Thank you to those who contributed to my thread.

ArmandoR1b
11-14-2016, 01:07 AM
Your post doesn't make any sense. Y-DNA does not recombine and is not affected by autosomal DNA and autosomal DNA is just a recombination of the parents autosomal DNA. Y-DNA does not affect autosomal DNA either.

ArmandoR1b
11-14-2016, 01:23 AM
Off the top of my head a good example that I can think of that directly refutes the misguided idea that Y-DNA and autosomal DNA correlate is the large number of people with only British Isles ancestry that are DF27 but have no Southern European autosomal DNA results in Family Finder but yet DF27 is very common in Iberia where everyone gets very high amounts of Southern European autosomal DNA results and many of them don't get any British Isles DNA. DF27 is also very common in Latin America and most of them also get very high amounts of Southern European autosomal DNA but many of those people don't get any British Isles DNA. You are most likely DF27 and you don't get any Southern European autosomal DNA in myOrigins.

sktibo
11-14-2016, 04:04 AM
So then would you say it's just more likely that MyOrigins is just wacky then Armando? Bad algorithm or something?

JMcB
11-14-2016, 04:11 AM
I'm afraid I don't really see any connection but I'll give you my results nevertheless. The My Origins results are: 100% European, 75% British Isles, 19% Southern European & 6% Eastern European and my Haplogroup is I-M253.

As can be expected, my known ancestry is primarily from the British Isles. Although, I do have some Italian and German ancestry through my paternal Grandfather who was 1/2 Irish, 1/4 Italian & 1/4 German. There is also a French line that dates back to the 1700s

sktibo
11-14-2016, 04:58 AM
I'm afraid I don't really see any connection but I'll give you my results nevertheless. The My Origins results are: 100% European, 75% British Isles, 19% Southern European & 6% Eastern European and my Haplogroup is I-M253.

As can be expected, my known ancestry is primarily from the British Isles. Although, I do have some Italian and German ancestry through my paternal Grandfather who was 1/2 Irish, 1/4 Italian & 1/4 German. There is also a French line that dates back to the 1700s

Your results definitely conflict with my hypothesis, thank you for sharing them

ArmandoR1b
11-14-2016, 01:33 PM
So then would you say it's just more likely that MyOrigins is just wacky then Armando? Bad algorithm or something?

The cause of myOrigins being wacky is a problem due to methodology, too few populations, and too few samples for many of the populations that were included. Simply compare the methodologies and populations between myOrigins and 23andme.

Here is the methodology and the populations for myOrigins - https://www.familytreedna.com/learn/user-guide/family-finder-myftdna/myorigins-methodology/

Here is the methodology for 23andme - https://www.23andme.com/ancestry-composition-guide/

You have to login to your 23andme account and go to the Ancestry Composition page to see the populations and sample sizes used.

It is scientifically impossible for the Y-DNA to affect the autosomal DNA which is what myOrigins uses.

AnnieD
11-15-2016, 05:47 AM
Edit: My comments below are in relation to replying on JohnHowellsTryfro's FTDNA results thread regarding my father's similar Y DNA and My Origins results.
____________________________________


The cause of myOrigins being wacky is a problem due to methodology, too few populations, and too few samples for many of the populations that were included. Simply compare the methodologies and populations between myOrigins and 23andme.

Love the technical gene-geek speak - "wacky" results. :lol: You must have been responding directly to my level of understanding. He He! I saw your comments on the FTDNA conference thread regarding the tiny sample sizes FTDNA used for My Origins. It doesn't even look like they visited more than 1 village for some of the sample sizes. I'll stop complaining about my intra-family dichotomy until the new & improved version is released. ;)


It is scientifically impossible for the Y-DNA to affect the autosomal DNA which is what myOrigins uses.

This makes sense, but I naively wondered if FTDNA knew that they had an imprecise methodology due to currently limited sample sized & therefore put an interpretative "spin" on the final results based on Y DNA origin assumptions. However, this should be beyond the time-frame of the My Origins results if truly representing roughly 1k years back. Perhaps another clue that your assumptions are spot-on is generally wide & rounded AC % at FTDNA in comparison to 23andMe, i.e. FTDNA = 85% this, 5% that, vs. 23andMe requires a calculator to check that you add up to 100% human - 61.5 British / Irish here & tiny 0.3% Finnish there. :D (but I add up to 101% at FTDNA. He he!)

I had not seen many parent-child FTDNA results for western Euros & wondered if our results were especially "wacky" given nearly 50% population in parent vs. 0% inheritance in child. Curious to see the new release.

Any opinions on whether Eurogenes K36 is similar time-frame to FTDNA My Origins? :confused:I've seen it reported variously that in spite of main function as chromosome painting tool it shows more ancient origins (but I would assume not along the lines of the new "Ancient Origins" offering by FTDNA).

wombatofthenorth
11-17-2016, 07:23 AM
Just for further evidence that they don't try to game the system if they see a Y-result for someone (and really would not make sense at all since it would just as often lead them astray), my dad has a pretty non-Baltic (non-Finnish and Siberian) Y-DNA (still a puzzle exactly where/when/how we got it), R-L20, and yet gets 88% Eastern European and 12 % Finnish and Siberian. So basically 100% of his ancestry composition assigned there is against his Y-DNA haplogroup so pretty strong evidence they don't do even the tiniest biasing at all.

AnnieD
11-17-2016, 02:55 PM
Just for further evidence that they don't try to game the system if they see a Y-result for someone (and really would not make sense at all since it would just as often lead them astray), my dad has a pretty non-Baltic (non-Finnish and Siberian) Y-DNA (still a puzzle exactly where/when/how we got it), R-L20, and yet gets 88% Eastern European and 12 % Finnish and Siberian. So basically 100% of his ancestry composition assigned there is against his Y-DNA haplogroup so pretty strong evidence they don't do even the tiniest biasing at all.

That definitely settles the matter for me! :thumb: OTOH ... your results as an East European-Baltic have looked consistently distinct, e.g. much higher Baltic component or Hunter-Gatherer % in Gedmatch calculators than other Euros so FTDNA test would not need to put an "interpretive spin" on your results. :P

However, now I'm right back where I started from - test methodology ? of how does a parent show 47% (call it 50% for ease of calculating) whereas child shows 0%? Have their been any other parent-child results at FTDNA that show this subregional variation of this degree? Granted, this is not as extreme or interesting a genetic problem as intercontinental would be. And 47%? Why not call it 50% if algorithm is struggling with low sample quantity & other issues? To use irregular % like 47% here & 48% there implies (at least in my untrained hobbyists eyes) a reliability factor. Can anyone explain the methodology used in layman's terms based on the link that ArmandoR1B provided below? :confused: Is there a Gedmatch calculator that shows parallel time-frame & hence can be compared to FTDNA's method? This is the kind of incongruity that makes me ? whether commercial DNA tests are a waste of time & $ for everyday hobbyists in regard to ethnicity or ancestry origins. Perhaps better to use as family finding / matching tools for now. :grouphug: