Identification of pathogenic missense mutations using protein stability predictors

Abstract

Attempts at using protein structures to identify disease-causing mutations have been dominated by the idea that most pathogenic mutations are disruptive at a structural level. Therefore, computational stability predictors, which assess whether a mutation is likely to be stabilising or destabilising to protein structure, have been commonly used when evaluating new candidate disease variants, despite not having been developed specifically for this purpose. We therefore tested 12 different stability predictors for their ability to discriminate between pathogenic and putatively benign missense variants. We find that one method, FoldX, considerably outperforms all others in the identification of disease variants. Moreover, we demonstrate that employing absolute energy change scores improves performance of nearly all predictors. Importantly, however, we observe that the utility of computational stability predictors is highly heterogeneous across different proteins, and that they are all are inferior to the best performing variant effect predictors for identifying pathogenic mutations. We suggest that this is largely due to alternate molecular mechanisms other than protein destabilisation underlying many pathogenic mutations. Thus, better ways of incorporating protein structural information and molecular mechanisms into computational variant effect predictors will be required for improved disease variant prioritisation.

Loss-of-function, gain-of-function and dominant-negative mutations have profoundly different effects on protein structure: implications for variant effect prediction

“…Our previous study on the identification of pathogenic mutations with protein stability predictors considered only monomeric proteins, in particular because many of the predictors we tested work only for single polypeptide chains 17 . As the dataset used in this study was derived from both monomeric and complex structures, we first investigated the impact of using full structures.…”

“…While these methods were not specifically designed for the identification of pathogenic variants, they are routinely used when evaluating candidate mutations [10][11][12][13] in order to identify those that are damaging to protein structure and will thus cause a loss of function (LOF) [14][15][16] . Alternatively, increased protein stabilisation can also be associated with disease, and it has been shown that using the absolute ΔΔG values results in higher accuracy when identifying disease mutations, although this may also be due to predictor inability to correctly distinguish the direction of the effect 17,18 . Interestingly, a recent study found that stability predictors performed much better in the identification of pathogenic missense mutations in genes associated with haploinsufficiency 19 , supporting the utility of stability predictors for identifying LOF mutations.…”

Protein structural features predict responsiveness to pharmacological chaperone treatment for three lysosomal storage disorders

“…One mechanism by which a missense mutation can exert a pathogenic effect is by destabilizing the protein in which it is located, leading to deficiency of the protein, not just its enzymatic activity (1)(2)(3)(4)(5). A smaller amount of the protein is then available to carry out its function, particularly since the destabilized protein is more likely to be cleared by quality control machinery (6)(7)(8).…”

Understanding protein structural changes for oncogenic missense variants

“…This is consistent with the observations from Figures 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, and 17, which do not show any consistent changes in the overall protein structures upon mutation, precluding any further statistical analysis. While we did not find overall principles that can be used to classify WT and oncogenic variables in this dataset, the judicious use of 3D structure prediction methods remains a valuable tool to understand oncogenic mutations further, as demonstrated previously [11,12,32,33,34,35]. We recognize that comparing these results with those that could be obtained from neutral non-pathogenic variants could provide more details on the problem and perhaps could help in differentiating the structural effects in different types of variants.…”