Let's Talk About HLA Nomenclature and Trans-species Variation

@John_Harshman has waited patiently to discuss trans-species variation and HLA nomenclature with me, after I put it on hold a while ago (John Harshman: Bottlenecks and Trans-Species Variation).

He has argued that there is trans-species variation for more than just 4 alleles at some HLA loci (locations/genes). However, I’m not sure. It seemed that variation I thought was a different loci, @John_Harshman was convinces was actually at the same locus. In the discussion, I was getting pretty confused by HLA nomenclature, because it seemed that @John_Harshman understood it quite differently than me.

This website is helpful:

So @John_Harshman can you please clarify specifically, with references, the evidence for more than 4 lineages of trans-species variation at a single loci? I honestly want to get this straight, and I think that clarifying the HLA nomenclature (which is complex but also tightly regulated) will help us do this.

1 Like

Here’s where we left off:

Note that these are all alleles of a single gene, HLA-DRB1. Oddly, HLA-DRB3 seems to be nested within it. Perhaps a gene duplication?


Let’s start by remembering that “loci” is plural and “locus” is singular. At this point I barely remember the relevant literature.


You should note that only humans, chimps, and macaques were assayed. I suspect that adding more apes, particularly gorillas and orangutans, would produce more shared alleles.

(I’ve been admonished to consolidate replies; it would help if I could post without waiting for approval. This is after all a thread designed for me specifically.)

1 Like

Incidentally, here is Yasukochi & Satta 2014. Fig. 3 is the main relevant part.


So your contention is that Group A and Group B are the same locus (correct), that I treated them as separate loci (did I?), and that if you add the trans-species variation from Group A and B it amounts to more than 4 lineages?

I have no idea whether you did or didn’t. I only point out what the tree shows.

Yes, and of course the variation that’s not within either group. It’s the whole tree you must consider. That’s the only point.

1 Like

So here is Fig 3 with the two cases of trans-species variation I can see. There are certainly other other HLA linages, but no other’s I’m seeing where the chimp sequence is within a human cluster. So taking the whole tree into account, I only count 2 alleles. How many do you count?


It is also possible that sampling chimps and macaques more would also identify more lineages. But let’s focus on the data we have already, not future data that might test the hypothesis more completely.

Six. Note that all of group A is sister to a chimp allele. That’s one.

HLA-DRB1*07 is sister to a chimp allele. But *09 and *04 are successive outgroups to that pair, and must therefore also be inherited diversity. That’s three more.

Your second cluster is another one. And finally, there’s *01:02:01, sister to all the other alleles.

Total of 6.

I see the issue. There is disagreement in how to define trans-species variation from this tree. I’m not sure how to resolve that disagreement.

Regardless, that figure alone soundly thrashes the hypothesis that we are not related to apes and monkeys. I’m not sure how you would explain that figure to a layperson, but it would be fun to watch.

1 Like

You could try believing the person who deals professionally with trees, character optimization, and relevant issues. Or you could look at it analytically. The node with a chimp allele on one side and a human on the other is, we agree, the point at which chimps and humans diverge. Nodes earlier (topologically) must predate that divergence. Therefore human alleles descending from those nodes must be inherited from primate ancestors, even if they aren’t directly accompanied by chimp alleles.


Yes, that’s true. Except for the possibility of convergent evolution here, but that’s another issue.

First off, is there any other data to look at that shows more than 4 alleles, or is this the only figure relevant?

Second, I’m open to be taught about this, but I’m having a hard time following your logic from words alone. You might need to annotate the image with the four lineages you are seeing.

I don’t know.

Sure. The red arrows point to alleles that must pre-date the chimp-human split, not because of any evolutionary rate assumptions but because of the topology of the tree.

HLA.pdf (1.2 MB)

1 Like

Whoops, I see that the arrow for Group A is misplaced. Move it down (well, left) by one node.