Application of Machine Learning Techniques to Predict Binding Affinity for Drug Targets: A Study of Cyclin-Dependent Kinase 2

BACKGROUND

The elucidation of the structure of cyclin-dependent kinase 2 (CDK2) made it possible to develop targeted scoring functions for virtual screening aimed to identify new inhibitors for this enzyme. CDK2 is a protein target for the development of drugs intended to modulate cellcycle progression and control. Such drugs have potential anticancer activities.

OBJECTIVE

Our goal here is to review recent applications of machine learning methods to predict ligand- binding affinity for protein targets. To assess the predictive performance of classical scoring functions and targeted scoring functions, we focused our analysis on CDK2 structures.

METHODS

We have experimental structural data for hundreds of binary complexes of CDK2 with different ligands, many of them with inhibition constant information. We investigate here computational methods to calculate the binding affinity of CDK2 through classical scoring functions and machine- learning models.

RESULTS

Analysis of the predictive performance of classical scoring functions available in docking programs such as Molegro Virtual Docker, AutoDock4, and Autodock Vina indicated that these methods failed to predict binding affinity with significant correlation with experimental data. Targeted scoring functions developed through supervised machine learning techniques showed a significant correlation with experimental data.

CONCLUSION

Here, we described the application of supervised machine learning techniques to generate a scoring function to predict binding affinity. Machine learning models showed superior predictive performance when compared with classical scoring functions. Analysis of the computational models obtained through machine learning could capture essential structural features responsible for binding affinity against CDK2.

On the mechanical characteristics of graphene nanosheets: a fully nonlinear modified Morse model

In this paper, the mechanical properties of graphene nanosheets are evaluated based on the nonlinear modified Morse model. The interatomic interactions including stretching and bending of the covalent bonds between carbon atoms, are replaced by nonlinear extensional and torsional spring-like elements. The finite element method is implemented to analyze the model under different loading conditions and linear characteristics of the graphene structure including the Young's modulus, surface modulus, shear modulus and Poisson's ratio are evaluated for various geometries and chirality where these properties are shown to be size and aspect ratio dependent. It is also found that when the dimensions of the sheets are greater than a certain threshold, the structure behaves quasi-isotropically and the directional elastic moduli become close to each other by a relative difference no more than 1%. Using the nonlinear stress-strain curve, the yielding point and ultimate stress and strains of the graphene sheet are also evaluated. The results of this study are compared with available experimental data and previous numerical simulations, where good agreement is achieved.

Generalization of the elastic network model for the study of large conformational changes in biomolecules

The elastic network (EN) is a prime model that describes the long-time dynamics of biomolecules. However, the use of harmonic potentials renders this model insufficient for studying large conformational changes of proteins (e.g. stretching of proteins, folding and thermal unfolding). Here, we extend the capabilities of the EN model by using a harmonic approximation described by Lennard-Jones (LJ) interactions for far contacts and native contacts obtained from the standard overlap criterion as in the case of Gō-like models. While our model is validated against the EN model by reproducing the equilibrium properties for a number of proteins, we also show that the model is suitable for the study of large conformation changes by providing various examples. In particular, this is illustrated on the basis of pulling simulations that predict with high accuracy the experimental data on the rupture force of the studied proteins. Furthermore, in the case of DDFLN4 protein, our pulling simulations highlight the advantages of our model with respect to Gō-like approaches, where the latter fail to reproduce previous results obtained by all-atom simulations that predict an additional characteristic peak for this protein. In addition, folding simulations of small peptides yield different folding times for α-helix and β-hairpin, in agreement with experiment, in this way providing further opportunities for the application of our model in studying large conformational changes of proteins. In contrast to the EN model, our model is suitable for both normal mode analysis and molecular dynamics simulation. We anticipate that the proposed model will find applications in a broad range of problems in biology, including, among others, protein folding and thermal unfolding.

Mechanical and vibrational characterization of amyloid-like HET-s nanosheets based on the skewed plate theory

Pathological amyloidogenic prion proteins have a toxic effect on functional cells in the human cerebrum because of poor degradability and the tendency to accumulate in an uncontrolled manner under physiological conditions. HET-s, a fungal prion protein, is known to undergo conformational variations from fibrillar to nanosheet structures during a change from low to high pH conditions. It has been said that this conformational change can lead to self-propagation by nucleating on the lateral surface of singlet fibrils. Efforts have been made toward the mechanical characterization of fibrillar amyloids, but a global understanding of amyloid-like HET-s nanosheet structures is lacking. In this study, we analyzed the mechanical and vibrational characteristics of the skewed HET-s nanosheet structures that developed under neutral pH conditions by performing various molecular dynamics simulations. By applying the skewed plate theory to HET-s nanosheets for various length scales with numerous pores inside the structures, we found that the skewed HET-s nanosheet structure has mechanical properties comparable to those of previously reported biological film materials and nanomaterials. Considering the inherent characteristics of structural stability, our observation provides valuable and detailed structural information on skewed amyloid-like HET-s nanosheets.