Knowledge-based Scoring Function Research Articles

Functional group based Ligand binding affinity scoring function at atomic environmental level.

Use of knowledge based scoring function (KBSF) for virtual screening and molecular docking has become an established method for drug discovery. Lack of a precise and reliable free energy function that describes several interactions including water-mediated atomic interaction between amino-acid residues and ligand makes distance based statistical measure as the only alternative. Till now all the distance based scoring functions in KBSF arena use atom singularity concept, which neglects the environmental effect of the atom under consideration. We have developed a novel knowledge-based statistical energy function for protein-ligand complexes which takes atomic environment in to account hence functional group as a singular entity. The proposed knowledge based scoring function is fast, simple to construct, easy to use and moreover it tackle the existing problem of handling molecular orientation in active site pocket. We have designed and used Functional group based Ligand retrieval (FBLR) system which can identify and detect the orientation of functional groups in ligand. This decoy searching was used to build the above KBSF to quantify the activity and affinity of high resolution protein-ligand complexes. We have proposed the probable use of these decoys in molecular build-up as a de-novo drug designing approach. We have also discussed the possible use of the said KSBF in pharmacophore fragment detection and pseudo center based fragment alignment procedure.

An iterative knowledge-based scoring function for protein-protein recognition.

Using an efficient iterative method, we have developed a distance-dependent knowledge-based scoring function to predict protein-protein interactions. The function, referred to as ITScore-PP, was derived using the crystal structures of a training set of 851 protein-protein dimeric complexes containing true biological interfaces. The key idea of the iterative method for deriving ITScore-PP is to improve the interatomic pair potentials by iteration, until the pair potentials can distinguish true binding modes from decoy modes for the protein-protein complexes in the training set. The iterative method circumvents the challenging reference state problem in deriving knowledge-based potentials. The derived scoring function was used to evaluate the ligand orientations generated by ZDOCK 2.1 and the native ligand structures on a diverse set of 91 protein-protein complexes. For the bound test cases, ITScore-PP yielded a success rate of 98.9% if the top 10 ranked orientations were considered. For the more realistic unbound test cases, the corresponding success rate was 40.7%. Furthermore, for faster orientational sampling purpose, several residue-level knowledge-based scoring functions were also derived following the similar iterative procedure. Among them, the scoring function that uses the side-chain center of mass (SCM) to represent a residue, referred to as ITScore-PP(SCM), showed the best performance and yielded success rates of 71.4% and 30.8% for the bound and unbound cases, respectively, when the top 10 orientations were considered. ITScore-PP was further tested using two other published protein-protein docking decoy sets, the ZDOCK decoy set and the RosettaDock decoy set. In addition to binding mode prediction, the binding scores predicted by ITScore-PP also correlated well with the experimentally determined binding affinities, yielding a correlation coefficient of R = 0.71 on a test set of 74 protein-protein complexes with known affinities. ITScore-PP is computationally efficient. The average run time for ITScore-PP was about 0.03 second per orientation (including optimization) on a personal computer with 3.2 GHz Pentium IV CPU and 3.0 GB RAM. The computational speed of ITScore-PP(SCM) is about an order of magnitude faster than that of ITScore-PP. ITScore-PP and/or ITScore-PP(SCM) can be combined with efficient protein docking software to study protein-protein recognition.

DrugScoreRNAKnowledge-Based Scoring Function To Predict RNA−Ligand Interactions

There is growing interest in RNA as a drug target due to its widespread involvement in biological processes. To exploit the power of structure-based drug-design approaches, novel scoring and docking tools need to be developed that can efficiently and reliably predict binding modes and binding affinities of RNA ligands. We report for the first time the development of a knowledge-based scoring function to predict RNA-ligand interactions (DrugScoreRNA). Based on the formalism of the DrugScore approach, distance-dependent pair potentials are derived from 670 crystallographically determined nucleic acid-ligand and -protein complexes. These potentials display quantitative differences compared to those of DrugScore (derived from protein-ligand complexes) and DrugScoreCSD (derived from small-molecule crystal data). When used as an objective function for docking 31 RNA-ligand complexes, DrugScoreRNA generates "good" binding geometries (rmsd (root mean-square deviation) < 2 A) in 42% of all cases on the first scoring rank. This is an improvement of 44% to 120% when compared to DrugScore, DrugScoreCSD, and an RNA-adapted AutoDock scoring function. Encouragingly, good docking results are also obtained for a subset of 20 NMR structures not contained in the knowledge-base to derive the potentials. This clearly demonstrates the robustness of the potentials. Binding free energy landscapes generated by DrugScoreRNA show a pronounced funnel shape in almost 3/4 of all cases, indicating the reduced steepness of the knowledge-based potentials. Docking with DrugScoreRNA can thus be expected to converge fast to the global minimum. Finally, binding affinities were predicted for 15 RNA-ligand complexes with DrugScoreRNA. A fair correlation between experimental and computed values is found (RS = 0.61), which suffices to distinguish weak from strong binders, as is required in virtual screening applications. DrugScoreRNA again shows superior predictive power when compared to DrugScore, DrugScoreCSD, and an RNA-adapted AutoDock scoring function.

An iterative knowledge-based scoring function to predict protein-ligand interactions: I. Derivation of interaction potentials.

Using a novel iterative method, we have developed a knowledge-based scoring function (ITScore) to predict protein-ligand interactions. The pair potentials for ITScore were derived from a training set of 786 protein-ligand complex structures in the Protein Data Bank. Twenty-six atom types were used based on the atom type category of the SYBYL software. The iterative method circumvents the long-standing reference state problem in the derivation of knowledge-based scoring functions. The basic idea is to improve pair potentials by iteration until they correctly discriminate experimentally determined binding modes from decoy ligand poses for the ligand-protein complexes in the training set. The iterative method is efficient and normally converges within 20 iterative steps. The scoring function based on the derived potentials was tested on a diverse set of 140 protein-ligand complexes for affinity prediction, yielding a high correlation coefficient of 0.74. Because ITScore uses SYBYL-defined atom types, this scoring function is easy to use for molecular files prepared by SYBYL or converted by software such as BABEL.

Substrate specificity in HIV‐1 protease by a biased sequence search method

Drug resistance in HIV-1 protease can also occasionally confer a change in the substrate specificity. Through the use of computational techniques, a relationship can be determined between the substrate sequence and three-dimensional structure of HIV-1 protease, and be utilized to predict substrate specificity. In this study, we introduce a biased sequence search threading (BSST) methodology to analyze the preferences of substrate positions and correlations between them that might also identify which positions within known substrates can likely tolerate sequence variability and which cannot. The potential sequence space was efficiently explored using a low-resolution knowledge-based scoring function. The low-energy substrate sequences generated by the biased search are correlated with the natural substrates. Octameric sequences were predicted using the probabilities of residue positions in the sequences generated by BSST in three ways: considering each position in the substrate independently, considering pairwise interdependency, and considering triple-wise interdependency. The prediction of octameric sequences using the triple-wise conditional probabilities produces the most accurate results, reproducing most of the sequences for five of the nine natural substrates and implying that there is a complex interdependence between the different substrate residue positions. This likely reflects that HIV-1 protease recognizes the overall shape of the substrate more than its specific sequence.

New Scoring Functions for Virtual Screening from Molecular Dynamics Simulations with a Quantum‐Refined Force‐Field (QRFF‐MD). Application to Cyclin‐Dependent Kinase 2.

AbstractChemInform is a weekly Abstracting Service, delivering concise information at a glance that was extracted from about 200 leading journals. To access a ChemInform Abstract, please click on HTML or PDF.

A knowledge-based scoring function based on residue triplets for protein structure prediction.

One of the general paradigms for ab initio protein structure prediction involves sampling the conformational space such that a large set of decoy (candidate) structures are generated and then selecting native-like conformations from those decoys using various scoring functions. In this study, based on a physical/geometric approach first suggested by Banavar and colleagues, we formulate a knowledge-based scoring function, which uses the radii of curvature formed among triplets of residues in a protein conformation. By analyzing its performance on various decoy sets, we determine a good set of parameters—the distance cutoff and the number of distance bins—to use for configuring such a function. Furthermore, we investigate the effect of using various approaches for compiling the prior distribution on the performance of the knowledge-based function. Possible extensions to the current form of the residue triplet scoring function are discussed.

M-Score: A Knowledge-Based Potential Scoring Function Accounting for Protein Atom Mobility

A knowledge-based potential scoring function, named M-Score, has been developed based upon 2331 high-resolution crystal structures of protein-ligand complexes. M-Score considers the mobility of protein atoms, describing the location of each protein atom by a Gaussian distribution instead of a fixed position based upon the isotropic B-factors. This leads to an increase in the number of atom-pairs in the construction of knowledge-based potentials and a smoothing effect on the pairwise distribution functions. M-Score was validated using 896 complexes which were not included in the 2331 data set and whose experimentally determined binding affinities were available. The overall linear correlation coefficient (r) between the calculated scores and experimentally determined binding affinities (pKi or pKd) for these 896 complexes is -0.49. Evaluation of M-Score against 17 protein families showed that we obtained good to excellent correlations for six protein families, modest correlations for four protein families, and poor correlations for the remaining seven protein families.

New Scoring Functions for Virtual Screening from Molecular Dynamics Simulations with a Quantum-Refined Force-Field (QRFF-MD). Application to Cyclin-Dependent Kinase 2

A recently introduced new methodology based on ultrashort (50-100 ps) molecular dynamics simulations with a quantum-refined force-field (QRFF-MD) is here evaluated in its ability both to predict protein-ligand binding affinities and to discriminate active compounds from inactive ones. Physically based scoring functions are derived from this approach, and their performance is compared to that of several standard knowledge-based scoring functions. About 40 inhibitors of cyclin-dependent kinase 2 (CDK2) representing a broad chemical diversity were considered. The QRFF-MD method achieves a correlation coefficient, R(2), of 0.55, which is significantly better than that obtained by a number of traditional approaches in virtual screening but only slightly better than that obtained by consensus scoring (R(2) = 0.50). Compounds from the Available Chemical Directory, along with the known active compounds, were docked into the ATP binding site of CDK2 using the program Glide, and the 650 ligands from the top scored poses were considered for a QRFF-MD analysis. Combined with structural information extracted from the simulations, the QRFF-MD methodology results in similar enrichment of known actives compared to consensus scoring. Moreover, a new scoring function is introduced that combines a QRFF-MD based scoring function with consensus scoring, which results in substantial improvement on the enrichment profile.

DrugScoreCSDKnowledge-Based Scoring Function Derived from Small Molecule Crystal Data with Superior Recognition Rate of Near-Native Ligand Poses and Better Affinity Prediction

Following the formalism used for the development of the knowledge-based scoring function DrugScore, new distance-dependent pair potentials are obtained from nonbonded interactions in small organic molecule crystal packings. Compared to potentials derived from protein-ligand complexes, the better resolved small molecule structures provide relevant contact data in a more balanced distribution of atom types and produce potentials of superior statistical significance and more detailed shape. Applied to recognizing binding geometries of ligands docked into proteins, this new scoring function (DrugScore(CSD)) ranks the crystal structures of 100 protein-ligand complexes best among up to 100 generated decoy geometries in 77% of all cases. Accepting root-mean-square deviations (rmsd) of up to 2 angstroms from the native pose as well-docked solutions, a correct binding mode is found in 87% of the cases. This translates into an improvement of the new scoring function of 57% with respect to the retrieval of the crystal structure and 20% with respect to the identification of a well-docked ligand pose compared to the original Protein Data Bank-based DrugScore. In the analysis of decoy geometries of cross-docking studies, DrugScore(CSD) shows equivalent or increased performance compared to the original PDB-based DrugScore. Furthermore, DrugScore(CSD) predicts binding affinities convincingly. Reducing the set of docking solutions to examples that deviate increasingly from the native pose results in a loss of performance of DrugScore(CSD). This indicates that a necessary prerequisite to successfully resolving the scoring problem with a more discriminative scoring function is the generation of highly accurate ligand poses, which approximate the native pose to below 1 angstroms rmsd, in a docking run.

Predicting Protein‐Ligand Binding Affinities Using Novel Geometrical Descriptors and Machine‐Learning Methods.

Inspired by the concept of knowledge-based scoring functions, a new quantitative structure−activity relationship (QSAR) approach is introduced for scoring protein−ligand interactions. This approach considers that the strength of ligand binding is correlated with the nature of specific ligand/binding site atom pairs in a distance-dependent manner. In this technique, atom pair occurrence and distance-dependent atom pair features are used to generate an interaction score. Scoring and pattern recognition results obtained using Kernel PLS (partial least squares) modeling and a genetic algorithm-based feature selection method are discussed.

Predicting protein-ligand binding affinities using novel geometrical descriptors and machine-learning methods.

Inspired by the concept of knowledge-based scoring functions, a new quantitative structure-activity relationship (QSAR) approach is introduced for scoring protein-ligand interactions. This approach considers that the strength of ligand binding is correlated with the nature of specific ligand/binding site atom pairs in a distance-dependent manner. In this technique, atom pair occurrence and distance-dependent atom pair features are used to generate an interaction score. Scoring and pattern recognition results obtained using Kernel PLS (partial least squares) modeling and a genetic algorithm-based feature selection method are discussed.

BHB: a simple knowledge-based scoring function to improve the efficiency of database screening.

A new knowledge-based scoring function was developed in this work to facilitate the rapid ranking of ligands in databases. The acronym of the method is BHB based on the descriptors it utilizes: buriedness, hydrogen bonding, and binding energy. Receptor buriedness is a measure of how well molecules occupy the binding pocket in comparison to known high-affinity ligands or, alternatively, whether they have contact with identified residues in the pocket. The possibility of hydrogen bond formation is checked for selected residues that are recognized as being important in the binding of known ligands. The approximate binding energy is calculated from the thermodynamic cycle using the optimized bound and free solvent conformations of the ligand-receptor system. The information necessary for the scoring function can ideally be gleaned from the 3D structure of the receptor-ligand complex. Alternatively, the descriptors can be derived from the 3D structure of the unbound receptor, provided this receptor has a known ligand that binds to the given site with nanomolar activity. We show that the new scoring functions provide up to 12 times improvement in enrichment compared to the popular commercial docking program GOLD.

SMall Molecule Growth 2001 (SMoG2001): an improved knowledge-based scoring function for protein-ligand interactions.

Computational lead design procedures require fast and accurate scoring functions to rank millions of generated virtual ligands for protein targets. In this article, we present an improved version of the SMoG scoring function, called SMoG2001. This function is based on a knowledge-based approach-that is, the free energy parameters are derived from the observed frequencies of atom-atom contacts in the database of three-dimensional structures of protein-ligand complexes via a procedure based on statistical mechanics. We obtained the statistics from the set of 725 complexes. SMoG2001 reproduces the experimental binding constants of the majority of 119 complexes of the testing set with good accuracy. On similar testing sets, SMoG2001 performs better than two other widely used scoring functions, PMF and SCORE1(LUDI), and comparably to DrugScore. SMoG2001 poorly predicts the affinities of ligands interacting via quantum mechanical forces with metal ions and ligands that are large and flexible. We attribute significant improvement in accuracy over previous versions of the SMoG scoring function to a better description of the reference state-that is, the state of no interactions.

From knowledge-based potentials to combinatorial lead design in silico.

Computational methods are becoming increasingly used in the drug discovery process. In this Account, we review a novel computational method for lead discovery. This method, called CombiSMoG for "combinatorial small molecule growth", is based on two components: a fast and accurate knowledge-based scoring function used to predict binding affinities of protein-ligand complexes, and a Monte Carlo combinatorial growth algorithm that generates large numbers of low-free-energy ligands in the binding site of a protein. We illustrate the advantages of the method by describing its application in the design of picomolar inhibitors for human carbonic anhydrase.

Identifying native-like protein structures using physics-based potentials.

As the field of structural genomics matures, new methods will be required that can accurately and rapidly distinguish reliable structure predictions from those that are more dubious. We present a method based on the CHARMM gas phase implicit hydrogen force field in conjunction with a generalized Born implicit solvation term that allows one to make such discrimination. We begin by analyzing pairs of threaded structures from the EMBL database, and find that it is possible to identify the misfolded structures with over 90% accuracy. Further, we find that misfolded states are generally favored by the solvation term due to the mispairing of favorable intramolecular ionic contacts. We also examine 29 sets of 29 misfolded globin sequences from Levitt's "Decoys 'R' Us" database generated using a sequence homology-based method. Again, we find that discrimination is possible with approximately 90% accuracy. Also, even in these less distorted structures, mispairing of ionic contacts results in a more favorable solvation energy for misfolded states. This is also found to be the case for collapsed, partially folded conformations of CspA and protein G taken from folding free energy calculations. We also find that the inclusion of the generalized Born solvation term, in postprocess energy evaluation, improves the correlation between structural similarity and energy in the globin database. This significantly improves the reliability of the hypothesis that more energetically favorable structures are also more similar to the native conformation. Additionally, we examine seven extensive collections of misfolded structures created by Park and Levitt using a four-state reduced model also contained in the "Decoys 'R' Us" database. Results from these large databases confirm those obtained in the EMBL and misfolded globin databases concerning predictive accuracy, the energetic advantage of misfolded proteins regarding the solvation component, and the improved correlation between energy and structural similarity due to implicit solvation. Z-scores computed for these databases are improved by including the generalized Born implicit solvation term, and are found to be comparable to trained and knowledge-based scoring functions. Finally, we briefly explore the dynamic behavior of a misfolded protein relative to properly folded conformations. We demonstrate that the misfolded conformation diverges quickly from its initial structure while the properly folded states remain stable. Proteins in this study are shown to be more stable than their misfolded counterparts and readily identified based on energetic as well as dynamic criteria. In summary, we demonstrate the utility of physics-based force fields in identifying native-like conformations in a variety of preconstructed structural databases. The details of this discrimination are shown to be dependent on the construction of the structural database.

Effect of ligand volume correction on PMF scoring

Recently, a knowledge-based scoring function has been introduced that estimates the protein-binding affinity based on the 3D structure of a protein–ligand complex (J Med Chem 1999, 42, 791). A ligand volume correction factor has been proposed and applied to filter out intraligand interactions in this simplified potential approach. Here we evaluate the effect of the ligand volume correction on the predictive power of the PMF scoring function. It is found that the effect of the ligand volume correction is significant on the derived potentials and large on the overall score. However, the effect of the ligand correction on the predictive power of the scoring function appears to be smaller. For a test set containing serine proteases the predictive power of the PMF scoring function does not change with the introduction of the volume correction. For a test set of metalloprotease complexes, the predictive power of the PMF scoring function improves only slightly when the volume correction is applied. For five test sets comprising a total of 225 diverse protein ligand complexes taken from the Brookhaven Protein Data Bank it is found, however, that the introduction of the ligand volume correction consistently improves the correlation between the PMF scores and the measured binding affinities. The effect of the correction factor on docking/scoring experiments is also analyzed using a test set of 61 biphenyl inhibitor-stromelysin complexes. © 2001 John Wiley & Sons, Inc. J Comput Chem 22: 418–425, 2001

Knowledge-based scoring function to predict protein-ligand interactions

The development and validation of a new knowledge-based scoring function (DrugScore) to describe the binding geometry of ligands in proteins is presented. It discriminates efficiently between well-docked ligand binding modes (root-mean-square deviation <2.0 Å with respect to a crystallographically determined reference complex) and those largely deviating from the native structure, e.g. generated by computer docking programs. Structural information is extracted from crystallographically determined protein-ligand complexes using ReLiBase and converted into distance-dependent pair-preferences and solvent-accessible surface (SAS) dependent singlet preferences for protein and ligand atoms. Definition of an appropriate reference state and accounting for inaccuracies inherently present in experimental data is required to achieve good predictive power. The sum of the pair preferences and the singlet preferences is calculated based on the 3D structure of protein-ligand binding modes generated by docking tools. For two test sets of 91 and 68 protein-ligand complexes, taken from the Protein Data Bank (PDB), the calculated score recognizes poses generated by FlexX deviating <2 Å from the crystal structure on rank 1 in three quarters of all possible cases. Compared to FlexX, this is a substantial improvement. For ligand geometries generated by DOCK, DrugScore is superior to the “chemical scoring” implemented into this tool, while comparable results are obtained using the “energy scoring” in DOCK. None of the presently known scoring functions achieves comparable power to extract binding modes in agreement with experiment. It is fast to compute, regards implicitly solvation and entropy contributions and produces correctly the geometry of directional interactions. Small deviations in the 3D structure are tolerated and, since only contacts to non-hydrogen atoms are regarded, it is independent from assumptions of protonation states.

Ab Initio fold prediction of small helical proteins using distance geometry and knowledge-based scoring functions

The problem of protein tertiary structure prediction from primary sequence can be separated into two subproblems: generation of a library of possible folds and specification of a best fold given the library. A distance geometry procedure based on random pairwise metrization with good sampling properties was used to generate a library of 500 possible structures for each of 11 small helical proteins. The input to distance geometry consisted of sets of restraints to enforce predicted helical secondary structure and a generic range of 5 to 11 Å between predicted contact residues on all pairs of helices. For each of the 11 targets, the resulting library contained structures with low RMSD versus the native structure. Near-native sampling was enhanced by at least three orders of magnitude compared to a random sampling of compact folds. All library members were scored with a combination of an all-atom distance-dependent function, a residue pair-potential, and a hydrophobicity function. In six of the 11 cases, the best-ranking fold was considered to be near native. Each library was also reduced to a final ab initio prediction via consensus distance geometry performed over the 50 best-ranking structures from the full set of 500. The consensus results were of generally higher quality, yielding six predictions within 6.5 Å of the native fold. These favorable predictions corresponded to those for which the correlation between the RMSD and the scoring function were highest. The advantage of the reported methodology is its extreme simplicity and potential for including other types of structural restraints.

Ab initio protein structure prediction using a combined hierarchical approach

As part of the third Critical Assessment of Structure Prediction meeting (CASP3), we predict the three-dimensional structures for 13 proteins using a hierarchical approach. First, all possible compact conformations of a protein sequence are enumerated using a highly simplified tetrahedral lattice model. We select a large subset of these conformations using a lattice-based scoring function and build detailed all-atom models incorporating predicted secondary structure. A combined all-atom knowledge-based scoring function is then used to select three smaller subsets from these all-atom models. Finally, a consensus-based distance geometry procedure is used to generate the best conformations from each of the all-atom subsets. With this method, we are able to predict the global topology/shape for all or a large part of the sequence for six out of the thirteen proteins. For two other proteins, the topology/shape for shorter fragments are predicted. This represents a marked improvement in ab initio prediction since CASP was first instigated in 1994.