[Newsbot]

Title: Twins get some 'mystifying' results when they put 5 ancestry DNA kits to the test
Category: Health
Posted By: DrCaleb
Date: 2019-01-18 07:08:06
Canadian

[BeaverFever]

I�ve done 2 of these tests, they�re fun but people need to take these results with a grain od salt. �Race�, �ethnicity� �nationality� and national borders are alll artificial, man-made concepts. And throughout history people have been intermarrying with people from other groups migrating to other lands and/or converting to other ethnicities. So of course these things can�t be identified with DNA. That doesn�t mean they�re total bunk either, it just means there are limits to what it can trell you. It can tell you that the particular combination of markers on your DNA are more commonly found in certain places and it will suggest that place as your recent ancestral homeland but there are many reasons that could be wrong.

[DrCaleb]

The fact that identical twins got different results from the same test means their accuracy is

[BeaverFever]

Butterfly effect. Small differences compound into large ones when plugged into algorithms. It’s the “ghost in the machine”. But note that while the twin results were not identical, they were still very similar.

[DrCaleb]

BeaverFever BeaverFever:

Butterfly effect. Small differences compound into large ones when plugged into algorithms.

Exactly. Shit.

If the algorithms weren't shit, they would produce the same result every time. Small differences would be seen as errors, and ignored.

[raydan]

You never know for sure who your ancestors really are, even though you might think you do. Something I heard about a long time ago is how much a new milkman could change the genetics in a neighbourhood.

[Freakinoldguy]

Who cares what they say.

Genealogy websites like Ancestory.com and 23andme are nothing more than DNA collecting receptacles for the government and police forces. In the past 3 months I've seen two people arrested decades after they committed their crimes because they took the DNA test which matched them up to the DNA from the crime scene.

That in itself would be fine if that's where it ended. But, if you think that's all they'll use this DNA fingerprint for I've got a very trendy New York bridge to sell.

[herbie]

They just take your money and make it up in their heads. Maybe use your last name to determine what element is over 50% and spinning wheels for the rest.
And nobody's gonna do shit about it if that's 100% true.

[BeaverFever]

DrCaleb DrCaleb:

BeaverFever BeaverFever:

Butterfly effect. Small differences compound into large ones when plugged into algorithms.

Exactly. Shit.

If the algorithms weren't shit, they would produce the same result every time. Small differences would be seen as errors, and ignored.

No, it’s not an error. The twins only share 99.6% of their DNA. Therefore their results were not identical. Go plug the number “100” into a complex mathematicial formula. Then go back and do it again with the number “99.6”. You’ll get a very different result. This is the butterfly effect, so named after weather models that varied drastically when numbers inputted in the calculation were rounded by a few decimal places by a lazy researcher, The difference from rounding the decimal places was metaphorically described as being the meteorological equivalent of the flap of a butterfly’s wings yet it resulted in a completely different weather pattern emerging every time.

https://en.m.wikipedia.org/wiki/Butterfly_effect


And how would the DNA company see an “error?” Each twins DNA was examined independently witthe computer doesn’t know

Go plug in the numbers

[Sunnyways]

The results are fairly accurate at a continental level but people want more than that for their money.

[DrCaleb]

BeaverFever BeaverFever:

DrCaleb DrCaleb:

BeaverFever BeaverFever:

Butterfly effect. Small differences compound into large ones when plugged into algorithms.

Exactly. Shit.

If the algorithms weren't shit, they would produce the same result every time. Small differences would be seen as errors, and ignored.

No, it’s not an error. The twins only share 99.6% of their DNA. Therefore their results were not identical.

'Error' is a changing limit, depending on what you are measuring. Getting 99.6% on a math test is functionally 100%, but 99.6% is not good enough to declare something a finding during a scientific study.

If a DNA test cannot get the same results given two nearly identical sets of DNA, then it is flawed. Would the test pick the same ethnic groups from their parents DNA that it picked for these girls? If not, then the test is not a test; it's accuracy no better than a palm reader or the readings of a tarot card.

And like FOG wrote, it's just a way for people to voluntarily give up very personal information about themselves, and have a little fun while doing it.

BeaverFever BeaverFever:

Go plug the number “100” into a complex mathematicial formula. Then go back and do it again with the number “99.6”. You’ll get a very different result. This is the butterfly effect, so named after weather models that varied drastically when numbers inputted in the calculation were rounded by a few decimal places by a lazy researcher, The difference from rounding the decimal places was metaphorically described as being the meteorological equivalent of the flap of a butterfly’s wings yet it resulted in a completely different weather pattern emerging every time.

https://en.m.wikipedia.org/wiki/Butterfly_effect


And how would the DNA company see an “error?” Each twins DNA was examined independently witthe computer doesn’t know

Go plug in the numbers

I took chaos theory in University. That's not how it works. There is nothing metaphoric about math. Meteorology, Chemistry and Biology are very different disciplines. There are many math formulas where 99.6 and 100 will give very similar answers, depending on the tolerance for error you are using. And a .4% difference can be a huge amount in others, again, depending on error rates.

The butterfly effect refers to how a slight difference can magnify over time (or do nothing at all), but .4% difference in DNA is always .4%. And they are always 99.6% identical. An algorithm that can't differentiate between small errors and important sequences is useless. The answers shouldn't diverge until the error rate is over 40%, not one hundreth of that.

The fact that even related siblings can get different results from the same tests calls their accuracy into question, but also because they have a very loose definition of 'ethnicity' means the tests can't be accurate.

https://news.nationalgeographic.com/201 ... s-science/

Once again, 'race' proves to be just a human concept, with little basis in science.

[BartSimpson]

I used Family Tree DNA and when my brother used it separate from my own it correctly matched him to me as my brother.

We've also discovered a first cousin we hadn't know about before.

Oh, and we discovered that we're Scottish and not Irish as we had been told when we were growing up. That explained a lot to us because we'd always observed a number of cultural differences between ourselves and our Irish friends.

[housewife]

Freakinoldguy Freakinoldguy:

Who cares what they say.

Genealogy websites like Ancestory.com and 23andme are nothing more than DNA collecting receptacles for the government and police forces. In the past 3 months I've seen two people arrested decades after they committed their crimes because they took the DNA test which matched them up to the DNA from the crime scene.

That in itself would be fine if that's where it ended. But, if you think that's all they'll use this DNA fingerprint for I've got a very trendy New York bridge to sell.

Actually it’s worse than that. Saw an interview with someone who used DNA from crime scene to build a family tree which eventually led to an arrest. The person arrested had never done a DNA test but cousins had.

[llama66]

That's how they caught the Golden State Killer last year. They are trying to use DNA to solve the Grim Sleeper case and the Zodiac murders.

[housewife]

Not saying it’s a bad thing that they are using these sites to solve these crimes. Just that there really isn’t any point to being paranoid about your DNA being known cause odds are some of it is already known. Between the family tree sites and the DNA sites there’s a lot of info floating around the internet for someone who knows what they are doing.