Question on 16189 mutation

[penguin]

Can someone tell me why having a mutation at 16189 prevents bidirectional sequencing?
(I'm also curious if anyone has anything interesting to say about the claim that having 16189 mutation causes a back mutation at 16192- supposedly the first evidence that local context can lead to a change? )

[cacio]

penguin:

I have not heard about backmutations at 16192, but I heard that with 16189C the length of the stretch can be very variable, and in fact it isn't usually reported in many scientific papers - and apparently testing companies don't check for it (that is, they don't count the exact number of C's). This was the only discrepancy between FTDNA and Argusbio in my HVRs, argusbio found 16193.1 C and ftdna didn't. I hear also that there can be heteroplasmy in the number of C's in the stretch, which would explain the discrepancy.

Another thing with the stretch is that this may cause mutations at the other end, ie 16183C also becomes likely, and possibly even 16182C (I have both). I have seen discussions on whether to interpret these as transversions (from A to C) or deletions (of the two regular A's), which is difficult to understand because of the variability in the number of C's in the area.

Nobody has been able to explain to me about the variability in the area, but it seems natural to assume it may have something to do with the length of the C stretch. I think 16189C creates a segment of 10 C's. With my other mutations, I should have 13 in a row.

cacio

[penguin]

This is the stretch from 16181 to 16200 in the rCRS
AAACCCCCTCCCCATGCTTA

So programs stop analyzing in the forward direction after 16189 if they get to a C in that position, which I guess is equivelent to saying the C's are not counted.

you're right, with c's at 182 and 183, an an insertion, you've got 13 in a row...

I don't have the mutation at 16189, but do have one (c to t) at 192, giving me only short chains of c's.

the back mutation of 192 i've been reading bout means if you;ve got a long stretch of c before 192, it seems to get rid of the T there (the 192 mutation is very old apparently and it's not clear how many times its gone back and forth)

anyway, only upshot is if you have the 189 mutation, that seems to imply you could end up being closer to people with a 192 mutaion than previously thought.

as you say the 189 is very common- i think in the new genographic paper -was it something like 17%- can't remember now.

good point that if you have those extra ones too- hard to know what kin of combination that was of insertions and deletions vs. some other transformaation.

you're also right about heteroplasmy (one of my favorite topics)- in the new genegraphic database, of all the heteroplasmy's found, that spot was one of the most frequent (data in one of the tables).

[cacio]

penguin:

thanks for the info, I should check the heteroplasmy part of the paper, as I assume it may apply to me as well.

I don't think I understand your explanation, simply because I don't know much about the topic. Is T the ancestral value of 16192? I don't think I've heard of backmutations of 16189, which ancestral state must be T.

I realized how common 16189C is because I belong to U1a, which is characterized by 16189 and 16249. But, given how frequent the 16189C mutation is, this combination pops out in other, unrelated haplogroups as well (F, B, M1, even L1c). Before FTDNA did checks on coding region mutations, they would systematically assign these sequences to U1a. I actually made a few people contact FTDNA to ask about their assignment, which was then regularly changed to their true haplogroup.

cacio

[penguin]

penguin:

I don't think I understand your explanation, simply because I don't know much about the topic. Is T the ancestral value of 16192? I don't think I've heard of backmutations of 16189, which ancestral state must be T.


cacio

I'm not sure if I know for sure if the ancestral value of 16192 is C or T - i've seen claims that would be consistent with either. I didn't mean to imply there's any back mutation of 16189; rather the claim was that a haplogroup that first had 16192T and later acquired 16189C would then revert back to 16192C, having been influenced by the 16189C (This imples ancestral is 16192C. On the other hand, I've seen at least one claim that 16192T goes back to Africa though I guess it still need not be ancestral). 16192 seems to be a hotspot-though i'd like to know if this is true for all haplgroups. i think it may be present more in yours than in mine actually.

yes, yours seems to be a haplogroup with so called homoplasy - or convergent evolution, as it would be called in other fields (same mutations from different lineages); I too got mine done before they did any coding region SNPs but mine seems pretty definitively T so its not as important as for U (and H and some others)

i so far still haven't sprung with argus for the full sequence- I had been set to order it at 345 and discovered they raised their price to 445 when I went to order it; i haven't quite gotten used to the idea yet.

[cacio]

penguin:

I quickly checked Ian Logan's webpage for L0 sequences, and didn't see 16192 among L0, so probably the ancestral value is C. However, mtdb displays 59 instances of the mutaiton, from all over the world. I didn't check what those sequences were, but given that they're from all over, they certainly belong to many different haplogroups, which confirms what you were saying about the variability of 16192. One of the papers cited in mtdb is a paper about sardinia, where the mutation defines a subgroup called U5b3. But even in that paper the mutation is found also in H.

Too bad that the price went up. I got mine when it was $350. Then a month ago I bought a kit for my dad, and of course it was already up. But remember, before argusbio came in, the only option was FTDNA at $1,000. So without argusbio we'd be still at those levels.

cacio

[penguin]

Check out table S-4 on the new genographic paper. it's useful for what we've been talking about- for any mutation of interest, we can see which haplogroups have that mutation and with what frequency! 16189 is all over the place, for instance, though in some haplogroups more than others it seems (they don't do the division to account for the different sizes of the different haplogroups though).

good point about how competition brought down the price for everyone.