Well, just taking rs3746544 as an example, the SNPedia page has this chart at the bottom of the right-hand column:
Notice there are only three colors there. Just eyeballing it, it seems most people have either the AA or AC genotype, with CC making up a pretty small minority.
But the fact that TT doesn't register at all on that chart and otherwise just isn't mentioned anywhere on the page at all makes me think I must be missing something. (Which wouldn't surprise me. It is genetics, after all. There are definitely tons of rabbit holes to fall down in that area of study.)
Someone else mentioned in a comment that it must be a mutation. (I assume they mean a mutation that happened recently enough -- few enough steps up my ancestral lineage -- that sources like SNPedia have just never seen it before.) And if I only had one example, and if it was only one allele and not on both strands, I might be inclined to agree.But I have... a lot of examples of that. Just doing some napkin math, I got a rate of about 16.5% "impossible" or undocumented genotypes. There's no way that for 16.5% of well-known/well-studied SNPs I happen to have a completely undocumented double-strand mutation, right?
(Just to explain my methodology for coming up with that 16.5% figure, I took all the SNP's listed on these two pages: one and two, and found all SNPs for which a) genotypes are documented on the page and b) I've got a genotype for that SNP in at least one of the data files I got from either 23andme or AncestryDNA (or both). That got me 139 examples. Of those, 23 of my SNP's didn't match any of the genotypes listed on the Eupedia pages. For a rate of 23/139=0.1654676...~=16.5% .)
(Side note: oddly of the 23 "mismatch" examples I mentioned, my genotype doesn't have a single allele in common with the documented possible alleles for the SNP. For example, I don't have any AT's where the documented alleles are AA, AC, and CC. My genes either match the documented alleles or have no alleles in common with the documented genotypes. Which seems even stranger.)
(Another side note: if 23andme and AncestryDNA didn't agree on the genotype, I'd be inclined to think it was an error on one of their parts, but I haven't found any specific SNPs where they disagree with each other yet.)
So, to get back to your main question:
Does “possible” mean “normal values for x% of the population?”
By "impossible", I mean I haven't been able to find any documentation of anyone else having the same genotype as I have for that particular SNP. And that makes me feel like I'm almost definitely not understanding something.
I kindof doubt that the genotypes I have for these mismatches are actually exceedingly rare or completely undocumented or anything. I think probably someone knows exactly why I'd be seeing the results I'm seeing (and it probably isn't tons of obscure mutations or anything.) So, honestly, "impossible" is probably bad wording.
If you were looking for a choir to preach to, I imagine this is about as choir a choir as one could find. Where this image really needs to go is on Xitter. Not that anyone here is likely to go register a Xitter account to do so.
You can take my terminal when you can pry it from my cold, dead, hands.
Any one-liner you put together, you can re-run trivially. You can rerun it with modifications trivially. You can wrap it in a for loop that runs it with different parameters trivially. You can stick it in a file and make a reusable Bash script. It's far easier to show someone else how you did it (just copy/paste the text of your terminal session) than dozens of screenshots of a point-and-click adventure (and not in a good way) GUI app. Bash commands are easier over SSH than GUI apps over RDP or VNC or whatever. You can't script a GUI app.
I seriously find myself wondering why someone would use a GUI for something they can do with a terminal. Learning curve is the only reason I can think of.
I guess we'll see if this post gets removed. If so, I guess maybe it's profanity. If not, maybe the AutoMod (or whoever built/configured it) thinks posts that short must be low-effort and remove them on that basis. Or it's straight up a glitch. Who knows.
My mother is constantly googling things and reading me the AI overview. And I know LLMs make shit up all the time, and I don't want AI hallucinations to infect my brain and slowly fuck up my worldview. So I always have to drop everything and go confirm the claims from the AI overview. And I've caught plenty of inaccuracies and hallucinations. (One I remember: she googled for when the East Wing of the White House was originally built and the AI overview told her the year of a major renovation, claiming it was the year it was built, but it had been built much earlier.)
Nope. I reply to bots quite frequently. I'm often the only commenter in the thread, but if I have something to say about what the bot posted, I'll comment.
Often my comments are to make fun of blockchain or LLMs. But if the Hacker News bot posts a link to a blog post on underwater beekeeping and I have an opinion, I'll post whatever I think adds to the conversation.
That sounds like exactly the sort of thing therapy is for. I'm no kind of expert, but it's very likely there's a lot of deeper things keeping you from developing achieving the kind of skills you're wanting. And it sounds very much like it's a problem in your life that's causing you a lot of anxiety and pain. I think if there's any way you can do talk therapy, that's the place to start.
Watching Amazon and Perplexity argue about AI agents making purchases on Amazon is like watching two discount birthday party clowns angrily honk bicycle horns at each other.
Wow. Huge topic. And it depends on a ton of things. And I definitely don't feel like I've got it all figured out myself.
If you're young and just for the first time having to manage your own affairs rather than depend on parents to help with that, then self-help kind of stuff might well be a fine place to start. (Just avoid Jordan Peterson.) If you're older and feel like you've had the time needed to develop those skills and still don't have them, it's likely there's something deeper going on that might benefit from therapy.
I personally cared for my ailing grandmother for a long time. And that shit's hard work, and takes a lot of time. In the process, I let a lot of things go by the wayside like yardwork, home repair, and organization. Now that she has passed, I find myself with a lot of remedial work to catch up on. I feel like I'm making progress. It's frustrating and slow, but it is progressing and that's the important part.
Well, just taking rs3746544 as an example, the SNPedia page has this chart at the bottom of the right-hand column:
Notice there are only three colors there. Just eyeballing it, it seems most people have either the AA or AC genotype, with CC making up a pretty small minority.
But the fact that TT doesn't register at all on that chart and otherwise just isn't mentioned anywhere on the page at all makes me think I must be missing something. (Which wouldn't surprise me. It is genetics, after all. There are definitely tons of rabbit holes to fall down in that area of study.)
Someone else mentioned in a comment that it must be a mutation. (I assume they mean a mutation that happened recently enough -- few enough steps up my ancestral lineage -- that sources like SNPedia have just never seen it before.) And if I only had one example, and if it was only one allele and not on both strands, I might be inclined to agree.But I have... a lot of examples of that. Just doing some napkin math, I got a rate of about 16.5% "impossible" or undocumented genotypes. There's no way that for 16.5% of well-known/well-studied SNPs I happen to have a completely undocumented double-strand mutation, right?
(Just to explain my methodology for coming up with that 16.5% figure, I took all the SNP's listed on these two pages: one and two, and found all SNPs for which a) genotypes are documented on the page and b) I've got a genotype for that SNP in at least one of the data files I got from either 23andme or AncestryDNA (or both). That got me 139 examples. Of those, 23 of my SNP's didn't match any of the genotypes listed on the Eupedia pages. For a rate of 23/139=0.1654676...~=16.5% .)
(Side note: oddly of the 23 "mismatch" examples I mentioned, my genotype doesn't have a single allele in common with the documented possible alleles for the SNP. For example, I don't have any AT's where the documented alleles are AA, AC, and CC. My genes either match the documented alleles or have no alleles in common with the documented genotypes. Which seems even stranger.)
(Another side note: if 23andme and AncestryDNA didn't agree on the genotype, I'd be inclined to think it was an error on one of their parts, but I haven't found any specific SNPs where they disagree with each other yet.)
So, to get back to your main question:
By "impossible", I mean I haven't been able to find any documentation of anyone else having the same genotype as I have for that particular SNP. And that makes me feel like I'm almost definitely not understanding something.
I kindof doubt that the genotypes I have for these mismatches are actually exceedingly rare or completely undocumented or anything. I think probably someone knows exactly why I'd be seeing the results I'm seeing (and it probably isn't tons of obscure mutations or anything.) So, honestly, "impossible" is probably bad wording.