Symmetry Principles in Optimization Problems: an application to Protein Stability Prediction

Pucci, Fabrizio; Bernaerts, Katrien V.; Teheux, Fabian; Gilis, Dimitri; Rooman, Marianne

doi:10.1016/j.ifacol.2015.05.068

Cited by 34 publications

(62 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When the prediction model is pre-established and not obtained through a black-box machine learning technique, it is possible to identify the terms in the model structure that are responsible for the symmetry breaking and appropriately correct them. This is exactly what we did in Pucci et al (2015), where the PoPMuSiC sym model, a symmetrized version of PoP-MuSiC v2.1, was presented.…”

Section: Designing Unbiased Prediction Modelssupporting

confidence: 77%

“…This implies that the predictors tend to be more accurate for destabilizing than for stabilizing mutations, which is a crucial issue given that the latter are the focus of protein design applications. This issue has been reported in a few investigations (Thiltgen and Goldstein (2012);Fariselli et al (2015); Pucci et al (2015)), but there is not yet a common, generally accepted, way to overcome it. Moreover, biases are not limited to this feature but can involve other characteristics such as the kind of protein or the type of wild type and mutant amino acids, since not all substitutions are sufficiently sampled in the training dataset.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Quantification of biases in predictions of protein stability changes upon mutations

Pucci

Bernaerts

Kwasigroch

et al. 2018

Preprint

View full text Add to dashboard Cite

Bioinformatics tools that predict protein stability changes upon point mutations have made a lot of progress in the last decades and have become accurate and fast enough to make computational mutagenesis experiments feasible, even on a proteome scale. Despite these achievements, they still suffer from important issues that must be solved to allow further improving their performances and utilizing them to deepen our insights into protein folding and stability mechanisms. One of these problems is their bias towards the learning datasets which, being dominated by destabilizing mutations, causes predictions to be better for destabilizing than for stabilizing mutations. We thoroughly analyzed the biases in the prediction of folding free energy changes upon point mutations (∆∆G 0 ) and proposed some unbiased solutions. We started by constructing a dataset S sym of experimentally measured ∆∆G 0 s with an equal number of stabilizing and destabilizing mutations, by collecting mutations for which the structure of both the wild type and mutant protein is available. On this balanced dataset, we assessed the performances of fifteen widely used ∆∆G 0 predictors. After the astonishing observation that almost all these methods are strongly biased towards destabilizing mutations, especially those that use blackbox machine learning, we proposed an elegant way to solve the bias issue by imposing physical symmetries under inverse mutations on the model structure, which we implemented in PoPMuSiC sym . This new predictor constitutes an efficient trade-off between accuracy and absence of biases. Some final considerations and suggestions for further improvement of the predictors are discussed.

show abstract

Section: Designing Unbiased Prediction Modelssupporting

confidence: 77%

Section: Introductionmentioning

confidence: 99%

Quantification of biases in predictions of protein stability changes upon mutations

Pucci

Bernaerts

Kwasigroch

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…With this objective in mind, we estimated with the PoPMuSiC sym algorithm [59,60] the change in folding free energy (∆∆G) for all single-site mutations in the non-redundant set D of protein X-ray structures representing the protein structurome, as described in Methods. In parallel, we considered the smaller ensembles of experimentally measured ∆∆G values and fitness scores.…”

Section: Resultsmentioning

confidence: 99%

“…We estimated the folding free energy changes (∆∆G) caused by all possible single-site mutations introduced in all collected structures, using the unbiased version of our in-house predictor PoPMuSiC, called PoPMuSiC sym [60,59]. The set of mutations so obtained is called M P oP .…”

Section: Large-scale In Silico Mutagenesis Experimentsmentioning

confidence: 99%

See 1 more Smart Citation

Large-scalein silicomutagenesis experiments reveal optimization of genetic code and codon usage for protein mutational robustness

Schwersensky

Rooman

Pucci

2020

Preprint

Self Cite

View full text Add to dashboard Cite

The question of how natural evolution acts on DNA and protein sequences to ensure mutational robustness and evolvability has been asked for decades without definitive answer. We tackled this issue through a structurome-scale computational investigation, in which we estimated the change in folding free energy upon all possible single-site mutations introduced in more than 20,000 protein structures. The validity of our results are supported by a very good agreement with experimental mutagenesis data. At the amino acid level, we found the protein surface to be more robust to mutations than the core, in a protein length-dependent manner. About 4% of all mutations were shown to be stabilizing, and a majority of mutations on the surface and in the core to be neutral and destabilizing, respectively. At the nucleobase level, single base substitutions were shown to yield on average less destabilizing amino acid mutations than multiple base substitutions. More precisely, the smallest average destabilization occurs for substitutions of base III in the codon, followed by base I, bases I+III, and base II. This ranking highly anticorrelates with the frequency of codon-anticodon mispairing, and suggests that the standard genetic code is optimized more to limit translation errors than the impact of random mutations. Moreover, the codon usage also appears to be optimized for minimizing the errors at the protein level, especially for surface residues that evolve faster and have therefore been under stronger selection, and for biased codons, suggesting that the codon usage bias also partly aims to optimize protein mutational robustness.

show abstract

Data set and fitting dependencies when estimating protein mutant stability: Toward simple, balanced, and interpretable models

Bæk

Kepp

2022

J Comput Chem

View full text Add to dashboard Cite

Accurate prediction of protein stability changes upon mutation (ΔΔG) is increasingly important to evolution studies, protein engineering, and screening of disease-causing gene variants but is challenged by biases in training data. We investigated 45 linear regression models trained on data sets that account systematically for destabilization bias and mutation-type bias B M . The models were externally validated on three test data sets probing different pathologies and for internal consistency (symmetry and neutrality). Model structure and performance substantially depended on training data and even fitting method. We developed two final models: SimBa-IB for typical natural mutations and SimBa-SYM for situations where stabilizing and destabilizing mutations occur to a similar extent. SimBa-SYM, despite is simplicity, is essentially non-biased (vs. the S sym data set) while still performing well for all data sets (R $ 0.46-0.54, MAE = 1.16-1.24 kcal/mol). The simple models provide advantage in terms of interpretability, use and future improvement, and are freely available on GitHub.

show abstract

Symmetry Principles in Optimization Problems: an application to Protein Stability Prediction

Cited by 34 publications

References 31 publications

Quantification of biases in predictions of protein stability changes upon mutations

Quantification of biases in predictions of protein stability changes upon mutations

Large-scalein silicomutagenesis experiments reveal optimization of genetic code and codon usage for protein mutational robustness

Data set and fitting dependencies when estimating protein mutant stability: Toward simple, balanced, and interpretable models

Contact Info

Product

Resources

About