iSEE: Interface Structure, Evolution and Energy-based machine learning predictor of binding affinity changes upon mutations

Geng, Cunliang; Vangone, Anna; Folkers, Gert E.; Li, Xue; Bonvin, Alexandre M. J. J.

doi:10.1101/331280

Cited by 6 publications

(16 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ELASPIC achieved a PCC of 0.75 and RMSE of 1.2638 kcal/mol on a subset of SKEMPI selected by applying a 90% sequence identity redundancy reduction, surpassing FoldX (PCC: 0.44) and BeAtMuSiC (PCC: 0.53) by a considerable margin. The RF algorithm has been most frequently used in ΔΔ G predictors (such as BindProf, iSEE, and MutaBind). BindProf constructs an interface binding profile score from an aligned ensemble of structurally similar interfaces, representing residue conservation.…”

Section: Predicting Binding Affinity Changes In Ppismentioning

confidence: 99%

“…Similarly, in the analysis of BindProf features with RF, FoldX ΔΔG and interface profile were the two most important features. The RF-based iSEE 65 predictor highlights the dominant role of residues conservation in ΔΔG prediction, followed by the changes of intermolecular electrostatic energy and changes of buried surface area. Taken together, these analyses underline the importance of intermolecular energies and residue conservation for estimating mutation effects on binding affinity of protein complexes.…”

Section: Important Featuresmentioning

confidence: 99%

See 1 more Smart Citation

Finding the ΔΔG spot: Are predictors of binding affinity changes upon mutations in protein–protein interactions ready for it?

Geng

Roel‐Touris

et al. 2019

WIREs Comput Mol Sci

Self Cite

View full text Add to dashboard Cite

Predicting the structure and thermodynamics of protein–protein interactions (PPIs) are key to a proper understanding and modulation of their function. Since experimental methods might not be able to catch up with the fast growth of genomic data, computational alternatives are therefore required. We present here a review dealing with various aspects of predicting binding affinity changes upon mutations (ΔΔG). We focus on predictors that consider three‐dimensional structure information to estimate the impact of mutations on the binding affinity of a protein–protein complex, excluding the rigorous free energy perturbation methods. Training and evaluation, ΔΔG databases, data selection, and existing ΔΔG predictors are specially emphasized. We also establish the parallel with scoring functions used in docking since those share many similar PPI features with ΔΔG predictors. The field has seen a common evolution of ΔΔG predictors and scoring functions over time, transforming from purely energetic functions to statistical energy‐based and further to machine learning‐based functions. As machine learning has come to age, limitations in terms of quantity, quality and variety of the available data become the bottlenecks for the future development of these computational methods. This can be alleviated by building infrastructures for data generation, collection and sharing. Further developments can be catalyzed by conducting community‐wide blind challenges for method assessment. This article is categorized under: Structure and Mechanism > Molecular Structures Structure and Mechanism > Computational Biochemistry and Biophysics Molecular and Statistical Mechanics > Molecular Interactions

show abstract

Section: Predicting Binding Affinity Changes In Ppismentioning

confidence: 99%

Section: Important Featuresmentioning

confidence: 99%

Finding the ΔΔG spot: Are predictors of binding affinity changes upon mutations in protein–protein interactions ready for it?

Geng

Roel‐Touris

et al. 2019

WIREs Comput Mol Sci

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the case of the complex of STXBP1 and syntaxin 1A, NT release is critically dependent on the structural details of the interaction. Structure-based predictors that consider PPIs were not developed and validated for in frame indels 27,28 . In this study, we demonstrate the pathogenicity of the R39dup variant based on the structure of the STXBP1-syntaxin 1A complex 29,30 .…”

Section: Main Textmentioning

confidence: 99%

Structural analysis ofde novoSTXBP1 mutation in complex with syntaxin 1A reveals a major alteration in the interaction interface in a child with developmental delay and spasticity

Banne

Falik‐Zaccai

Brielle

et al. 2019

Preprint

View full text Add to dashboard Cite

STXBP1, also known as Munc-18, is a master regulator of neurotransmitter release and synaptic function in the human brain through its direct interaction with syntaxin 1A. STXBP1 related disorders are well characterized and cover a diverse range of neurological and neurodevelopmental conditions. Through exome sequencing of a child with developmental delay, hypotonia and spasticity, we found a novel de novo insertion mutation of three nucleotides in the STXBP1 coding region, resulting in an additional arginine after position 39 (R39dup).Inconclusive results from state-of-the-art variant prediction tools mandated a structure-based approach using molecular dynamics (MD) simulations of the STXBP1-syntaxin 1A complex.Comparison of the interaction interfaces of the wild type and the R39dup complexes revealed a reduced interaction surface area in the mutant, leading to destabilization of the interaction. We applied the same MD methodology to 7 additional previously reported STXBP1 mutations. We find that the stability of the STXBP1-syntaxin 1A interface correlates with the reported clinical phenotypes. We illustrate a direct link between a patient's genetic variations and the observed clinical phenotype through protein structure, dynamics, and function.

show abstract

“…Hence, they suffer from poor generalizability and lead to low accuracy. With the increasing availability of large mutation databases, statistical learning algorithms have been proposed to capture the relations between a variety of energetic or structural features and the binding affinity of two biomolecules [13, 29, 43, 51]. Nevertheless, features used in these methods are hand-crafted, requiring extensive human labors and expert knowledge.…”

Section: Introductionmentioning

confidence: 99%

“…However, none of the existing methods can directly estimate the BSA score, or the change in BSA upon mutation. BSA is often computed from the ASA scores of individual proteins and the protein complex [13].…”

Section: Introductionmentioning

confidence: 99%

Mutation effect estimation on protein-protein interactions using deep contextualized representation learning

Zhou

Chen

et al. 2019

Preprint

View full text Add to dashboard Cite

The functional impact of protein mutations is reflected on the alteration of conformation and thermodynamics of protein-protein interactions (PPIs). Quantifying the changes of two interacting proteins upon mutations are commonly carried out by computational approaches. Hence, extensive research efforts have been put to the extraction of energetic or structural features on proteins, followed by statistical learning methods to estimate the effects of mutations to PPI properties. Nonetheless, such features require extensive human labors and expert knowledge to obtain, and have limited abilities to reflect point mutations. We present an end-to-end deep learning framework, , to estimate the effects of mutations on PPIs. incorporates a contextualized representation mechanism of amino acids to propagate the effects of a point mutation to surrounding amino acid representations, therefore amplifying the subtle change in a long protein sequence. On top of that, leverages a Siamese residual recurrent convolutional neural encoder to encode a wildtype protein pair and its mutation pair. Multiple-layer perceptron regressors are applied to the protein pair representations to predict the quantifiable changes of PPI properties upon mutations. Experimental evaluations show that outperforms various state-of-the-art systems on the change of binding affinity prediction and the buried surface area prediction. The software implementation is available at https://github.com/guangyu-zhou/MuPIPR

show abstract

iSEE: Interface Structure, Evolution and Energy-based machine learning predictor of binding affinity changes upon mutations

Cited by 6 publications

References 43 publications

Finding the ΔΔG spot: Are predictors of binding affinity changes upon mutations in protein–protein interactions ready for it?

Finding the ΔΔG spot: Are predictors of binding affinity changes upon mutations in protein–protein interactions ready for it?

Structural analysis ofde novoSTXBP1 mutation in complex with syntaxin 1A reveals a major alteration in the interaction interface in a child with developmental delay and spasticity

Mutation effect estimation on protein-protein interactions using deep contextualized representation learning

Contact Info

Product

Resources

About