Prediction and interpretation of deleterious coding variants in terms of protein structural stability

Abstract

The classification of human genetic variants into deleterious and neutral is a challenging issue, whose complexity is rooted in the large variety of biophysical mechanisms that can be responsible for disease conditions. For non-synonymous mutations in structured proteins, one of these is the protein stability change, which can lead to loss of protein structure or function. We developed a stability-driven knowledge-based classifier that uses protein structure, artificial neural networks and solvent accessibility-dependent combinations of statistical potentials to predict whether destabilizing or stabilizing mutations are disease-causing. Our predictor yields a balanced accuracy of 71% in cross validation. As expected, it has a very high positive predictive value of 89%: it predicts with high accuracy the subset of mutations that are deleterious because of stability issues, but is by construction unable of classifying variants that are deleterious for other reasons. Its combination with an evolutionary-based predictor increases the balanced accuracy up to 75%, and allowed predicting more than 1/4 of the variants with 95% positive predictive value. Our method, called SNPMuSiC, can be used with both experimental and modeled structures and compares favorably with other prediction tools on several independent test sets. It constitutes a step towards interpreting variant effects at the molecular scale. SNPMuSiC is freely available at https://soft.dezyme.com/ .

Figure 1

Schematic representation of: (a) the artificial neural network and (b) the probabilistic neural network used in the classification of the variants.

Figure 2

Probability density distributions of deleterious mutations (red curve) and neutral mutations (blue curve) for: (a) the solvent accessibility A (0–100%) of the mutated residues; (b) the change in volume ΔV (in ${\AA }^3$) for residues with A ≤ 60%; (c) the change in volume ΔV for residues with A > 60%; in our conventions, mutations of smaller into larger residues have a positive ΔV value.

Figure 3

Probability density distributions of deleterious mutations (red curve) and neutral mutations (blue curve) for the changes in folding free energy ΔΔW (in kcal/mol) computed with the following statistical potentials: (a) the distance potential $\mathrm{\Delta }{W}_{sd}$; (b) the distance potential $\mathrm{\Delta }{W}_{sds}$; (c) the torsion angle and solvent accessibility potential $\mathrm{\Delta }{W}_{sta}$; in our conventions, positive $\mathrm{\Delta }\mathrm{\Delta }W$ values correspond to destabilizing mutations.

Figure 4

Probability density distributions of deleterious mutations (red curve) and neutral mutations (blue curve) for (a) the change in folding free energy ΔΔG computed by PoPMuSiC (Eq. (3)) (in kcal/mol), (b) the Provean score^,, and (c) the pathogenicity index I computed by the ANN model (Eq. (8)).

Figure 5

Probability density distribution of deleterious mutations (red curve) and neutral mutations (blue curve) for the pathological index $\mathcal{J}$ (Eq. (11)) computed by SNPMuSiC. The distribution curves in the high confidence intervals, which lie from either side of the two vertical lines, are depicted on a white background.