Adam never met Eve - coalescence time for HLA-DRB1 alleles

Date: Thu Nov 02 2000 - 14:30:12 EST

  • Next message: glenn morton: "RE: Adam never met Eve - coalescence time for HLA-DRB1 alleles"

    The following 2 papers are relevant to this discussion:
    R.E. Hickson & R.L. Cann, "Mhc allelic diversity and modern human
    origins", J.Mol.Evol. 45 (1997), 589-98.
    T.F. Bergstrom, A. Josephson, H.A. Erlich & U. Gyllensten, "Recent
    origin of HLA-DRB1 alleles and implications for human evolution", Nature
    Genetics 18 (1998), 237-42.

    They show that the vast majority of DRB alleles are relatively young
    (order of magnitude: 250,000 years), consistent with other nuclear and
    with mtDNA data, and indicate a population size of about 10^4
    individuals for most of the last 1 million years. The previous
    coalescence estimates ignored the DRB groups, the concentration of the
    polymorphisms in exon 2, the coding sequences being under selection and
    other problems. Of course, coalescence time estimates are extremely

    Bergstrom et al.'s abstract:
    "The HLA class I and class II loci are the most highly polymorphic
    coding regions in the human genome. Based on the similarity of the
    coding sequences of alleles between species, it has been claimed that
    the HLA polymorphism is ancient and predates the separation of human
    (Homo) and chimpanzee (Pan), 4-7.4 Myr ago. Analysis of intron
    sequences, however, provides support for a more recent origin and for
    rapid generation of alleles at the HLA class II DRB1 locus. The human
    DRB1 alleles can be divided into groups (allelic lineages); most of
    these lineages have diverged from each other before the separation of
    Homo and Pan. Alleles within such a lineage, however, appear to be, on
    average, 250,000 years old, implying that the vast majority (greater
    than 90%) of the more than 135 contemporary human DRB1 alleles have been
    generated after the separation of Homo and Pan. The coalescence time of
    alleles within allelic lineages indicates that the effective population
    size (Ne) for early hominids (over the last 1 Myr) was approximately
    10^4 individuals, similar to the estimates based on other nuclear loci
    and mitochondrial DNA. With a single exception, the genetic mechanisms
    (gene conversion and point mutation) that have diversified the exon-2
    sequences do not appear to extend into the adjacent intron sequences.
    The part of exon 2 encoding the beta-sheet evolves in concert with the
    surrounding introns, while the alpha-helix appears to have been
    subjected to gene conversion-like events, suggesting that such exchange
    events are highly localized and occur over extremely short sequence

    Furthermore, they write:
    "The phylogenetic trees for primate exon-2 sequences... are consistent
    with the notion that most of the alleles at some class II loci, such as
    DRB1 and DPB1, may have a more recent origin [than previously believed]"
    ... "[S]tudies of human populations with a defined degree and time of
    isolation (for example South American Indians) have provided support for
    the view that new DRB1 alleles have been generated over the last
    10,000-20,000 years."

    "Given that coding sequences are subject to selection, they are
    problematic for determining the evolutionary relationships and age of
    alleles. Numerous studies have shown the polymorphism at HLA class I and
    II loci to be under selection. For instance, the codons for amino acids
    involved in binding of foreign peptides (antigen recognition site, ARS)
    show an excess of replacement changes relative to silent changes, while
    no such effect has been observed at the non-ARS codons, consistent with
    diversifying selection. Furthermore, the polymorphism at several class
    II loci, such as DRB1, is characterized by a 'patchwork' pattern of
    amino-acid motifs, indicating that the alleles may have been generated
    through sequence exchanges (that is, gene conversion-like events). This
    implies that adjacent coding sequences may not share the same
    evolutionary history and are therefore not suited to reconstructing
    evolutionary relationships among alleles."

    "The mean sequence difference among alleles within a lineage corresponds
    to an average age of 180,000-320,000 years (range based on the standard
    error). This implies that the vast majority of the more than 135
    contemporary HLA-DRB1 alleles have a very recent origin." ... "Because
    the ARS and non-ARS codons do not share the same evolutionary history,
    ... a realistic estimate of the age of alleles cannot be obtained using
    exon sequences."

    Hickson & Cann came to similar conclusions.

    Furthermore, the HLA are the proteins presenting foreign peptides to the
    immunity-generating system. If the parasites and other antigens humans
    and other primates had to fend off were similar, we should expect
    parallel evolution of their HLA epitope groups, which would simulate
    erroneously low rates of diversifying evolution, and therefore too high
    coalescence times.

    Peter Ruest <>

    This archive was generated by hypermail 2b29 : Thu Nov 02 2000 - 14:27:55 EST