PDA

View Full Version : 3% Irish Mom + 1% Irish Dad = 5% Irish Son, Possible?



Adison
03-13-2017, 01:47 PM
Thanks beforehand for your assistance.

Recently received results from AncestryDNA. The title indicates the question.

3% Irish Mom + 1% Irish Dad = 5% Irish Son, possible?

My results came in this way. Logically it doesn't seem to make sense. Recessive heritage genes would mean anything is possible, no?

I am guessing it's going to be 'margin of error situation'.

Thanks again.

03-13-2017, 03:18 PM
Thanks beforehand for your assistance.

Recently received results from AncestryDNA. The title indicates the question.

3% Irish Mom + 1% Irish Dad = 5% Irish Son, possible?

My results came in this way. Logically it doesn't seem to make sense. Recessive heritage genes would mean anything is possible, no?

I am guessing it's going to be 'margin of error situation'.

Thanks again.

If I understand Chromosome splitting correctly. Your son gets 50% from you and 50% from your wife.
But.. and here is the big but, It could be that your son would then get a bigger dose of irish % from one or more of the grand parents. hence getting more than you or your wife.

Adison
03-13-2017, 03:22 PM
If this is true, and a grand child could inherit more then his parents inherit, then anything would be possible re child's composition relative to parents. Correct? This seems illogical wacky.

For example, a child is 100% Irish. His parents have 0% Irish between them. However grand parents have enough Irish in them to skip a generation and compile 100%. Whew.

MitchellSince1893
03-13-2017, 03:34 PM
If this is true, and a grand child could inherit more then his parents inherit, then anything would be possible re child's composition relative to parents. Correct? This seems illogical wacky.

For example, a child is 100% Irish. His parents have 0% Irish between them. However grand parents have enough Irish in them to skip a generation and compile 100%. Whew.


If your mother was 3.4% and your father was 1.4% you could theoretically be 4.8%.

I had a similar question about my German ancestry a couple of years ago and here is what I came up with.

http://www.anthrogenica.com/showthread.php?2184-Understanding-my-Autosomal-results&p=112405&viewfull=1#post112405

Adison
03-13-2017, 03:39 PM
Mitchell, just read your post. Thanks.

Rounding... hmm.

MitchellSince1893
03-13-2017, 03:50 PM
Mitchell, just read your post. Thanks.

However, you get to the sum of your parents.

In this case I am exceeding the sum of my parents.

It could be due rounding. 3.4% rounded to 3%. 1.4% rounded to 1%. Sum 4.8% rounded to 5%

Adison
03-13-2017, 04:08 PM
I will have more to add when 23andme results come in. Thanks.

I would like to confirm, taking rounding out of the equation, a child cannot exceed the sum of his/her parents, correct?

Dewsloth
03-13-2017, 04:17 PM
I will have more to add when 23andme results come in. Thanks.

I would like to confirm, taking rounding out of the equation, a child cannot exceed the sum of his/her parents, correct?

Well no, but the calculators are not the ultimate truth in the matter. They may see the same information and group it differently.

I score higher ashkenazi than either of my parents. Dad: 9% Mom: 0% Me: 11%

http://www.anthrogenica.com/showthread.php?9695-Extreme-MyOrigins-Differences-Between-Father-and-Son&p=212837&viewfull=1#post212837

C J Wyatt III
03-13-2017, 04:28 PM
I will have more to add when 23andme results come in. Thanks.

I would like to confirm, taking rounding out of the equation, a child cannot exceed the sum of his/her parents, correct?

You are assuming a precision here that isn't. In looking at matching segments, we postulate things like a match has to have 500 SNPs and 7.0 cM minimum segment size. Obviously that is just a feeble attempt on our part to try and bring some sense to the data. I have seen cases where both parents share a common ancestor, but neither has enough DNA from that ancestor to meet our matching criteria when they compare with another descendent of that common ancestor. However the DNA from the two parents comes together and a match mysteriously appears with their child to that other person who shares the common ancestor.

I would say that admixture analysis theory is even less developed than matching theory. For now, I don't think your example is far enough off to worry about.

Jack Wyatt

geebee
03-14-2017, 04:23 PM
If you could take your parents' 3% and 1% literally, then no, you could not end up with 5%. The most you'd get would be the combined total of 4%.

However, it might be helpful to look at how Ancestry arrives at these figures. On the page that says "Ethnicity estimate for [Your Name]", there are links to a lot of useful bits of information. Two that you should look at are entitled "How is the ethnicity estimate determined?" and "How is the range calculated?"

The first of these tells you a little about the reference panels, which is nice to know; but more importantly, it has a link to a "White paper" which gives more of the scientific background. For your purposes, what is important to know is that all of your percentages are only estimates. And if you were to run the test 40 times with slightly different parameters, you might get 40 different estimates.

So that's actually what they do. They run the test 40 times. But each time, they "cut" your DNA into slightly different chunks. If you think about it, this is really essential. How can they know, in advance, where each segment begins and ends? If you're comparing two cousins, you can (mostly) see where runs of matching DNA begin and end, but when trying to determine ancestry, who are they comparing you to? Not individuals, but whole panels.

So what they're doing is dividing your DNA into somewhat arbitrary segments, and then trying to determine the likely ancestral origin of each segment. But since the segments are arbitrary, there's really no way to know for sure if the ancestry is really the same through the whole segment. (In other words, is it really a segment at all?)

By doing this 40 times, though, with different divisions, they arrive at 40 estimates for each ancestry that appears. Some ancestries almost certainly appear on some runs, but not often enough or with a high enough percentage to include them in the final result.

Your final percentage for any given ancestry that is among those that "made the cut" is actually the average of the 40 runs. Your range gives you your lowest to highest test results. So, for example, my largest component is "Europe West", with 44%. But, I actually had individual test results as low as 15% and as high as 71%.

So, your parents' 3% and 1% represent their average results. Your 5% also represents your average. What this tells you, among other things, is that at least one of these numbers can't be exactly right. Not taken literally, anyway.

What if Ancestry did a new set of 40 runs for each of you -- this time using even slightly different "sectioning" of your DNA? It's possible that your average would also be slightly different. I know that a couple of my numbers are different between v1 and v2. I don't think it's necessarily because of the different versions, so much as the new set of 40 runs.

For example, although my Europe West is the same 44% for both versions, my ranges are 15%-71% for v1, and 15%-72% for v2. My Great Britain result is actually different between the two: 27% for v1, and 25% for v2. The ranges are 0%-56% and 0%-53%, respectively.

So, as I said, don't take your numbers (or your parents' numbers) too literally. Just because they've given you the average for each test's results doesn't actually mean that's the best result. For any given ethnicity, the best result needn't even have been from the same test run. Maybe run number 17 just happened to result in a more accurate grouping for your Irish segments, even though it wasn't the same as your average, for example.

But don't just go by what I've said here. I don't always say things as well as I'd like (I can never use 10 words when 50 are available, for example). So read the white paper, and also look at the descriptions of each region -- especially both the "primarily located in" and "also found in" parts.