Table 3.
Species | MHC region | nalleles | Pi | dN/dS | dN-dS | SE | Z-score | P |
---|---|---|---|---|---|---|---|---|
Human | HLA-I exon 2 (α1) | 83 | 0.087 | 3.160 | 0.203 | 0.081 | 0.728 | 0.233 |
HLA-I exon 3 (α2) | 83 | 0.065 | 2.447 | 0.123 | 0.076 | 0.052 | 0.479 | |
HLA-I exon 2 & 3 (α1α2) | 83 | 0.076 | 2.809 | 0.161 | 0.054 | 1.565 | 0.059 | |
HLA-II DRA exon 2 (α1) | 1 | 0.000 | NA | 0.000 | 0.000 | |||
HLA-II DRB1 exon 2 (β1) | 29 | 0.062 | 1.922 | 0.117 | 0.086 | |||
HLA-II DR- A & B1 exon 2 (α1β1) | 29 | 0.033 | 1.750 | 0.050 | 0.046 | |||
Common buzzard | MHC-I exon 2 (α1) | 10 | 0.094 | 3.461 | 0.219 | 0.075 | -1.147 | 0.126 |
MHC-I exon 3 (α2) | 8 | 0.065 | 2.635 | 0.121 | 0.059 | -2.042 | 0.021 | |
MHC-I exon 2 & 3 (α1α2) | 10 | 0.079 | 3.089 | 0.166 | 0.042 | -0.143 | 0.443 | |
MHC-IIA exon 2 (α1) | 3 | 0.008 | NA | 0.035 | 0.036 | |||
MHC-IIB exon 2 (β1) | 8 | 0.105 | 9.156 | 0.367 | 0.105 | |||
MHC-II A & B exon 2 (α1β1) | 8 | 0.056 | 7.481 | 0.175 | 0.047 |
To demonstrate how a false conclusion can be reached when comparing only one set of peptide binding domains between MHC classes, we collected and analyzed well-curated alleles from humans and full alleles of common buzzards. Human class I HLA-A, B, C, and class II HLA-DRB1 alleles from the European population classified as "common" in the Common and Well-Documented allele catalog 3.0.0 [70] as well as HLA-DRA*01:01 from the monomorphic DRA locus were downloaded from IPD-IMGT/HLA Database version 3.51 [16] (Additional file Table S5). Common buzzard alleles with full peptide-binding grooves were retrieved from genomic data (GenBank accession #s: OL311287, OL311290, OL311292, OL311294, OL311304, OL311305, OP490259, OQ414190-OQ414202, OQ428163-OQ428174). All analyses for the strength of selection (dN and dS) were conducted in MEGA X [71] using human peptide binding residues from [72] and the Nei-Gojobori model [73] with 1000 bootstraps for variance estimation. The Z-score was calculated with the formula (dN-dSMHC-I—dN-dSMHC-II)/(SE2MHC-I + SE2MHC-II). We compared MHC-I exon 2 and 3 separately with MHC-IIB exon 2, and then combined against concatenated MHC-IIA & B exon 2 sequences. We tested if MHC-I had stronger selection than MHC-II in humans and if MHC-IIB had stronger selection than MHC-I in common buzzards. One-sided p-values were generated from the Z-scores. P-values show how comparing single exons between the classes does not represent comparisons across the full binding domains (italicized for emphasis; MHC-I: α1α2, MHC-II: α1β1). nalleles: number of alleles, Pi: nucleotide diversity, SE: standard error of the mean. Significant p-values are highlighted in bold