Fig. 6.

Prediction of the aggregation properties of β₂m and an example variant, F62A. (a) Intrinsic propensity for aggregation of wild-type β₂m (grey, left axis) as predicted by the Zyggregator algorithm modified to acidic conditions.²⁵ The difference in aggregation propensity between wild-type β₂m and its variant F62A (black, right axis). F62A is predicted by $Z_{\text{i}}^{\text{agg}}$ to have only a small local effect on the intrinsic aggregation propensity of β₂m. A schematic illustration of the observed effects of amino acid substitution on the apparent rate of fibril elongation taken from Fig. 3d is shown above. The effects are classified as in Fig. 3, with rates not significantly altered compared with wild-type β₂m (grey), 2- to 4-fold change (hashed) and > 4-fold change (black). (b) Comparison of the smoothed R₂ relaxation rates for wild-type β₂m and F62A β₂m. Plots of R₂ rates on a per-residue basis for all variants are shown in Fig. 8. (c) Comparison of the aggregation propensities of wild-type and the difference in aggregation propensities between wild-type and F62A β₂m obtained by modification of the Zyggregator algorithm by the measured R₂ values (${\tilde{Z}}_{\text{i}}^{\text{agg}}$).