Recursive analysis of the amino acid sequence within the peptide-binding site of DRB1 locus. (A) Lower panel: WebLogo visualization representing the contribution of single amino acids within the variable portion of the peptide-binding site of DRB1 locus. The x-axis indicates each variable position (as per IPD-IMGT-HLA reference). Letters represent each possible amino acid at each given position. Letter height illustrates the frequency of each amino acid in a healthy control population. Colors indicate the chemico-physical properties as per legend in panel B. Upper panel: stylized visualization of amino acids differentially distributed between HC and IAA/PNH cohorts. Letter height illustrates the frequency of each amino acid in the disease population. Markers indicate the OR resulting from the logistic regression analysis studying each amino acidic contribution in determining the phenotype (see supplemental Methods and supplemental Table 10). (B) Scatterplot showing the significance level of each variable amino acid in the peptide-binding site of DRB1 locus found enriched in IAA/PNH population compared with HC. Each dot represents the negative log₁₀ for the adjusted P value (q value) referring to each amino acid. The position on the x-axis indicates the position within the peptide-binding site according to IPD-IMGT-HLA reference. Only the amino acids presenting a q value <10-e4 are considered significant for this analysis and labeled in the figure. Alleles presenting the indicated amino acid at the given position are indicated in the boxes. Colors represent the chemico-physical properties as per legend. (C) Crystallographic structure showing the position of the 7 amino acids significantly enriched in the peptide-binding groove of DR molecules of IAA/PNH patients. This 3D structure has been visualized with the PyMOL program based on the structure of DRB1*15:01-DRA*01:01-myelin binding protein (PDB:1BX2) and HLA sequences retrieved from IPD-IMGT/HLA database v.3.40. Only the 7 amino acids identified in the previous analysis have been highlighted. (D) Binding site of DRB1*16:01-DRA*01:01 based on the homology model of DRB1*16:01. The structure of DRB1*16:01 has been superimposed to DRB1*15:01. Residues differing from the risk pattern seen in DRB1*15:01 are colored in red. (E) Binding site of DRB1*12:01-DRA*01:01 based on the homology model of DRB1*12:01. The structure of DRB1*12:01 has been superimposed to the structure of DRB1*15:01. Residues differing from the risk pattern seen in DRB1*15:01 are colored in red. aa, amino acid; HC, healthy control.