Canonical DNA questions

4:17 PM

This chat room has been created to discuss the scope structure and creation of canonicial DNA Q&S.

Assuming we agree on the need for canonical DNA questions, I'm proposing 3:

1. Dealing with ethnicity estimates -- understanding them, how accurate they are, what if they don't match what I know about my ancestry...

2. Interpreting relationship estimates -- what do they mean, what other possibilities exist.

Those first two are about nuts and bolts but the third is more about ethics and privacy.

3. I think I've identified a close relative -- how can I be sure? How should I proceed? Also, somebody I expected to show us as a close relative doesn't? What does this mean? How do I proceed?

Of course, there may be a better grouping of topics?

Once we've agreed structure and scope, they can be created as Community Wiki questions and worked on in parallel.

Cyn

4:41 PM

I am here but don't know how to use this feature.

While I agree about the need for a Q on DNA ethnicity, I'm not sure about the grouping of the DNA relatives. Because most questions that come up are about both. Ideally, we'll want to be able to point one question to one canonical question, only occasionally to two or (God forbid) more.

That's why I suggested:
1. My DNA results show a very close family match to someone I don't know, how should I proceed?

2. My close relative did not show up in my DNA matches as expected.

ColeValleyGirl

@Cyn You're doing fine -- if you clock to the left of a message you'll see a down arrow which allows you to tag message as a reply which should mean the person you're replying to gets an alart.

Alert even

Cyn

Alert?

ColeValleyGirl

Re grouping, my thought was there are a bunch of questions that aren't about close relatives but are about interpreting the relationship estimates and using that info to further paper research. Which is kind of mostly a nuts and bolt thing -- no emotional or privacy concerns.

The close relative present/absent has all sorts of emotional baggage that doesn't need to be addressed for straight estimatiion of the relationships -- so I'd see those questions pointed to a canonical question that referred to the 'nuts and bolts' material to explain what the relationships might be but majored on the harder bits.

@Cyn Possibly in your Stack 'inbox' or as a visible/audible alert on your desktop.

Cyn

I see. While yes we should have a FAQ or other method to point to for the basics, my reason for posting the Meta Q was because of the rash of questions that were all about close relatives that were new people or old people who didn't match as expected.

Another reason I'd like to see a can. post about close relative matching is because there are not a lot of choices and we could outline them pretty efficiently all at once. From parent to first cousin I'd say.

ColeValleyGirl

@Cyn I'd kind of like to avoid ending up with material duplicated in two canonical questions -- and we do get questions about the less-emotional analysis.

I don't think putting the basic stuff in a tag wiki would help -- various discussions in the past have concluded that won't work.

Cyn

4:55 PM

I think "how do I interpret my DNA match result?" is the overarching question and that it's different from "ohmygod I have a surprise close family match!"

ColeValleyGirl

@Cyn I agree about the overarching question, and that it's different from the surprise close family match (or absence of match, which I would put in the same question -- oh my god he's not my father!)

Maybe if we try to outline the content of each question and answer, we can see if the overlaps matter?

Cyn

Let me ask a basic question here first though...
By "canonical question" do you mean the following:
-One or more of us posts a question such as what we've been discussing, as a regular question on the site.
-Authorship is community.
-One or more of us answers it comprehensively.
-We allow other answers.
-The best answer is accepted (probably the one by whoever of us we designated to write it) and therefore appears on top.
-When new questions come in, we refer them there, either as duplicates or as part of an answer, whichever is appropriate.

"One or more of us" = one of us writes it or it's collaborative, but it's just one question per topic.

ColeValleyGirl

@Cyn Communitti wikie -- see genealogy.meta.stackexchange.com/a/3383/6485

Cyn

Nodding.

ColeValleyGirl

A standard question that is made community wiki (ditto the answer), so there's no reputation accruing and anyone can improve them. Hopefully a single collaboratively created (accepted) answer -- multiple answers run the risk of muddying the waters and if the answer is CW there'd have to be a really good reason reason to breed answers. One question per topic.

Yes, refer questions to them either as complete duplicates or part of the answer (if there really is a unique element to the question that ought not be handled by updateing the CW post).

Cyn

5:07 PM

"Ohmygod I got a SURPRISE DNA result!"

Answer:

First, check the cM numbers to see which relationships are possible.
(List all of them with cM ranges)
(Link to cM Match Project and brief explanation)

Second, look for other known matches on the DNA testing site to narrow things down.
(Explanations and examples)

Third, use the DNA info to draw other conclusions.
(Explanations and examples, like Y, mt, ethnicity, matches on both vs one side of parents/grandparents, etc)

Fourth, how to talk to the new relative or the old one you don't match as expected, how to ask family about family secret…

ColeValleyGirl

I did a DNA test and don't understand what it's telling me about matches.

What do segments and centimorgans mean?

How reliable are the estimates the website gives me for possible relationships? What other possibilites are there and how likely are they?

Why do different websites give me different suggestions/different numbers?

What do I do next?

That's an outline for the overarching question

Cyn

Yep.

That looks good.

I'd say perhaps an entirely separate one for adoption...I mean the first will cover a lot of it but adoptees also grab on to even tiny matches (I know because I'm on the receiving end of tons of them...anyone endogamous community member is). So it is a different frame of an answer.

ColeValleyGirl

Answer would cover each point in term -- terminology, relationship estimates (with pointer to the DNA painter relationship probability tool, why websites differ (looking at different SNPs, difference calculation bases especially for siblings -- there's one major outlier), then the stuff about fishing in wider pools, triangulation, follow the paper trail...

I agree re adoption. Altough we need to cover off the significance of match size in the overarching question as well?

Did we want a separate endogamy question -- although probably get the first two of three done first (and I know nothing about endogamy).

Cyn

Yeah, there's a big diff between close matches (sib/aunt), matches that are often in known family (first/second cousin), and more distant matches one doesn't generally know about even if there are no family secrets (3C+)

I would incorporate endogamy into all the questions as needed. I hate when essays completely ignore it (here's how to triangulate!) and I also when I'm shunted off to a specialty article.

ColeValleyGirl

Fair point -- but you'll probably have to add that in -- all I know is that endogmany means that 'objects in the rear view mirror may appear closer than they are'

Cyn

5:19 PM

LOL that's not a bad way to describe it. But yes, I'm quite familiar with endogamy.

Even the statement that endogamy is NOT an issue for this case (like with close family) is important, because some people often think it is and it causes confusion and can chase away people from end. communities.

ColeValleyGirl

I have to go and cook dinner (UK time here) but if you're agreeable, I'll work on the overarching question and answer offline and post them as a text attachment (see the Upload button next to send) for suggestions/corrections/additions etc. and you do the same for the Surprise question? I suppose the overarching one ought to cover the different types on DNA testing as well (makes a mental note).

Once we (and anyone else who joins in here) are reasonably happy the the stuff we've done is fit for the public, we can post it and ask one of the Mods to make it community wiki?

Cyn

Okay, sounds good. And yes, because some testing sites include mt or Y with the autosonomal (23andme does, FTDNA notes them if tested for separately, and Gedmatch allows you to add them...and note that both 23andme and Gedmatch give a Y for me, even though I am XX, because my dad tested)

That sounds like a plan!

and yeah different timing...I had a pause before going into the chat because I had to take my daughter to school. It's 9:30am here (Calif).

ColeValleyGirl

See you later

(much later :) )

Cyn

LOL see ya

Cyn

8:50 PM

I get an error when I try to upload the file. Here it is on Dropbox: dropbox.com/s/uxxw8io1pbmbub7/…

Transcript for

Canonical DNA questions