Predicting Protein Binding Research Articles

Experimental verification of structural alerts for the protein binding of sulfur-containing compounds

As often noted by Dr. Gilman Veith, a major barrier to advancing any model is defining its applicability domain. Sulfur-containing industrial organic chemicals can be grouped into several chemical classes including mercaptans (RSH), sulfides (RSR’), disulfides (RSSR’), sulfoxides (RS(=O)R’), sulfones (RS(=O)(=O)R’), sulfonates (ROS(=O)(=O)R’) and sulfates (ROS(=O)(=O)OR’). In silico expert systems that predict protein binding reactions from 2D structure sub-divide these chemical classes into a variety of chemical reactive mechanisms and reactions which have toxic consequences. Using the protein binding profilers in version 3.1 of the OECD QSAR Toolbox, a series of sulfur-containing chemicals were profiled for protein binding potential. From these results it was hypothesized which sulfur-containing chemicals would be reactive or non-reactive in an in chemico glutathione assay and whether if reactive they would exhibit toxicity in excess of baseline in the Tetrahymena pyriformis population growth impairment assay. Subsequently, these hypotheses were tested experimentally. The in chemico data show that the in silico profiler predictions were generally correct for all chemical categories, where testing was possible. Mercaptans could not be assessed for GSH reactivity because they react directly with the chromophore 5,5’-dithiobis-(2-nitrobenzoic acid). With some exceptions, the major being disulfides, the in vitro toxicity data supported the in chemico findings.

Predicting protein-binding RNA nucleotides using the feature-based removal of data redundancy and the interaction propensity of nucleotide triplets

Several learning approaches have been used to predict RNA-binding amino acids in a protein sequence, but there has been little attempt to predict protein-binding nucleotides in an RNA sequence. One of the reasons is that the differences between nucleotides in their interaction propensity are much smaller than those between amino acids. Another reason is that RNA exhibits less diverse sequence patterns than protein. Therefore, predicting protein-binding RNA nucleotides is much harder than predicting RNA-binding amino acids. We developed a new method that removes data redundancy in a training set of sequences based on their features. The new method constructs a larger and more informative training set than the standard redundancy removal method based on sequence similarity, and the constructed dataset is guaranteed to be redundancy-free. We computed the interaction propensity (IP) of nucleotide triplets by applying a new definition of IP to an extensive dataset of protein-RNA complexes, and developed a support vector machine (SVM) model to predict protein binding sites in RNA sequences. In a 5-fold cross-validation with 812 RNA sequences, the SVM model predicted protein-binding nucleotides with an accuracy of 86.4%, an F-measure of 84.8%, and a Matthews correlation coefficient of 0.66. With an independent dataset of 56 RNA sequences that were not used in training, the resulting accuracy was 68.1% with an F-measure of 71.7% and a Matthews correlation coefficient of 0.35. To the best of our knowledge, this is the first attempt to predict protein-binding RNA nucleotides in a given RNA sequence from the sequence data alone. The SVM model and datasets are freely available for academics at http://bclab.inha.ac.kr/primer.

Predicting permanent and transient protein-protein interfaces.

Protein-protein interactions (PPIs) are involved in diverse functions in a cell. To optimize functional roles of interactions, proteins interact with a spectrum of binding affinities. Interactions are conventionally classified into permanent and transient, where the former denotes tight binding between proteins that result in strong complexes, whereas the latter compose of relatively weak interactions that can dissociate after binding to regulate functional activity at specific time point. Knowing the type of interactions has significant implications for understanding the nature and function of PPIs. In this study, we constructed amino acid substitution models that capture mutation patterns at permanent and transient type of protein interfaces, which were found to be different with statistical significance. Using the substitution models, we developed a novel computational method that predicts permanent and transient protein binding interfaces (PBIs) in protein surfaces. Without knowledge of the interacting partner, the method uses a single query protein structure and a multiple sequence alignment of the sequence family. Using a large dataset of permanent and transient proteins, we show that our method, BindML+, performs very well in protein interface classification. A very high area under the curve (AUC) value of 0.957 was observed when predicted protein binding sites were classified. Remarkably, near prefect accuracy was achieved with an AUC of 0.991 when actual binding sites were classified. The developed method will be also useful for protein design of permanent and transient PBIs.

SCOWLP update: 3D classification of protein-protein, -peptide, -saccharide and -nucleic acid interactions, and structure-based binding inferences across folds

BackgroundProtein interactions are essential for coordinating cellular functions. Proteomic studies have already elucidated a huge amount of protein-protein interactions that require detailed functional analysis. Understanding the structural basis of each individual interaction through their structural determination is necessary, yet an unfeasible task. Therefore, computational tools able to predict protein binding regions and recognition modes are required to rationalize putative molecular functions for proteins. With this aim, we previously created SCOWLP, a structural classification of protein binding regions at protein family level, based on the information obtained from high-resolution 3D protein-protein and protein-peptide complexes.DescriptionWe present here a new version of SCOWLP that has been enhanced by the inclusion of protein-nucleic acid and protein-saccharide interactions. SCOWLP takes interfacial solvent into account for a detailed characterization of protein interactions. In addition, the binding regions obtained per protein family have been enriched by the inclusion of predicted binding regions, which have been inferred from structurally related proteins across all existing folds. These inferences might become very useful to suggest novel recognition regions and compare structurally similar interfaces from different families.ConclusionsThe updated SCOWLP has new functionalities that allow both, detection and comparison of protein regions recognizing different types of ligands, which include other proteins, peptides, nucleic acids and saccharides, within a solvated environment. Currently, SCOWLP allows the analysis of predicted protein binding regions based on structure-based inferences across fold space. These predictions may have a unique potential in assisting protein docking, in providing insights into protein interaction networks, and in guiding rational engineering of protein ligands. The newly designed SCOWLP web application has an improved user-friendly interface that facilitates its usage, and is available at http://www.scowlp.org.

Binding Site Prediction for Protein-Protein Interactions and Novel Motif Discovery using Re-occurring Polypeptide Sequences

BackgroundWhile there are many methods for predicting protein-protein interaction, very few can determine the specific site of interaction on each protein. Characterization of the specific sequence regions mediating interaction (binding sites) is crucial for an understanding of cellular pathways. Experimental methods often report false binding sites due to experimental limitations, while computational methods tend to require data which is not available at the proteome-scale. Here we present PIPE-Sites, a novel method of protein specific binding site prediction based on pairs of re-occurring polypeptide sequences, which have been previously shown to accurately predict protein-protein interactions. PIPE-Sites operates at high specificity and requires only the sequences of query proteins and a database of known binary interactions with no binding site data, making it applicable to binding site prediction at the proteome-scale.ResultsPIPE-Sites was evaluated using a dataset of 265 yeast and 423 human interacting proteins pairs with experimentally-determined binding sites. We found that PIPE-Sites predictions were closer to the confirmed binding site than those of two existing binding site prediction methods based on domain-domain interactions, when applied to the same dataset. Finally, we applied PIPE-Sites to two datasets of 2347 yeast and 14,438 human novel interacting protein pairs predicted to interact with high confidence. An analysis of the predicted interaction sites revealed a number of protein subsequences which are highly re-occurring in binding sites and which may represent novel binding motifs.ConclusionsPIPE-Sites is an accurate method for predicting protein binding sites and is applicable to the proteome-scale. Thus, PIPE-Sites could be useful for exhaustive analysis of protein binding patterns in whole proteomes as well as discovery of novel binding motifs. PIPE-Sites is available online at http://pipe-sites.cgmlab.org/.

INCB38579 is a novel, potent and selective histamine H4 receptor antagonist with activity in rodent models of inflammatory pain (117.28)

Abstract The histamine H4 receptor (H4) has been considered as an attractive therapeutic target for a number of diseases, such as allergy, asthma, atopic dermatitis and inflammatory pain. INCB38579 is a novel human and rodent active H4 receptor antagonist, with at least 50-fold selectivity over the human H1, H2 and H3 receptors and good pharmacokinetic properties in rats and mice. The compound potently inhibits histamine binding to and signaling through recombinant human and mouse H4 and blocks histamine-induced migration of human and mouse dendritic cells, as well as cell shape change and migration of human eosinophils. This novel H4 inhibitor is efficacious in blocking histamine-induced pruritus in mice and carrageenan-induced acute inflammatory pain in rats. Interestingly, when examined in a formalin pain model, INCB38579 could significantly reduce the relatively long lasting inflammation associated pain. A good correlation between the predicted protein binding adjusted potency from in vitro studies and the analgesic effect in vivo was observed. These results suggest that INCB38579 can serve as a useful tool for pharmacologic characterization of H4 and further supports the hypothesis that the H4 receptor is a target for the development of new therapeutics for various inflammatory diseases.

Prediction of protein binding regions

Identifying protein binding sites provides important clues to the function of a protein. Experimental methods to identify the binding sites such as determining the crystal structures of protein complexes are extremely laborious and expensive. Here, we present a computational technique called spatial aggregation propensity (SAP) based on molecular simulations to predict protein binding sites. We apply this technique to two model proteins, an IgG1 antibody and epidermal growth factor receptor (EGFR) and demonstrate that SAP predicts protein binding regions with very good accuracy. In the case of the IgG1 antibody, SAP accurately predicts binding regions with the Fc-receptor, protein-A, and protein-G. For EGFR, SAP accurately predicts binding regions with EGF, TGFα, and with another EGFR. The resolution of SAP is varied to obtain a detailed picture of these binding sites. We also show that some of these binding sites overlap with protein self-aggregation prone regions. We demonstrate how SAP analysis can be used to engineer the protein to remove unfavorable aggregation prone regions without disturbing protein binding regions. The SAP technique could be also used to predict the yet unknown binding sites of numerous proteins, thereby providing clues to their function.

The binding site distance test score: a robust method for the assessment of predicted protein binding sites

We propose a novel method for scoring the accuracy of protein binding site predictions-the Binding-site Distance Test (BDT) score. Recently, the Matthews Correlation Coefficient (MCC) has been used to evaluate binding site predictions, both by developers of new methods and by the assessors for the community-wide prediction experiment-CASP8. While being a rigorous scoring method, the MCC does not take into account the actual 3D location of the predicted residues from the observed binding site. Thus, an incorrectly predicted site that is nevertheless close to the observed binding site will obtain an identical score to the same number of non-binding residues predicted at random. The MCC is somewhat affected by the subjectivity of determining observed binding residues and the ambiguity of choosing distance cutoffs. By contrast the BDT method produces continuous scores ranging between 0 and 1, relating to the distance between the predicted and observed residues. Residues predicted close to the binding site will score higher than those more distant, providing a better reflection of the true accuracy of predictions. The CASP8 function predictions were evaluated using both the MCC and BDT methods and the scores were compared. The BDT was found to strongly correlate with the MCC scores while also being less susceptible to the subjectivity of defining binding residues. We therefore suggest that this new simple score is a potentially more robust method for future evaluations of protein-ligand binding site predictions. http://www.reading.ac.uk/bioinf/downloads/.

Where are the boundaries? Automated pocket detection for druggability studies

Computer-based prediction of protein druggability is an essential task in the drug development process. Early identification of disease modifying targets that can be modulated by low-molecular weight compounds can help to speed up and reduce costs in drug discovery. Recently, first methods have been presented performing a druggability estimation solely based on the 3D structure of the protein [1-3]. The essential first step for such methods is the identification of the active site. A multitude of methods exist for automated active site prediction [4-6]. However, most methods developed for automated docking procedures do not explicitly focus on the definition of the boundary of the active site. Since druggability estimates are based on structural descriptors of the active site, a precise description of the active site boundaries is vital for correct predictions. In this work, we present a method to predict protein binding pockets and split them into subpockets such that small molecules are mostly contained within one sub-pocket. The method is based on a novel strategy to geometrically detect narrow regions in pockets. For druggability predictions, such pocket descriptions result in more meaningful structural descriptors like active site surface or volume. Moreover, if several structures from one protein are known, sub-pockets can give hints about protein flexibility and induced fit conformational changes. Our method was evaluated on 718 proteins from the PDBbind [7] data set, as well as 5419 proteins from the scPDB [8] data set. Binding pockets are correctly predicted in 94% and 93% of the datasets. 38%, respectively 45% of the proteins from the two datasets contain pockets which can be divided into more than one sub-pocket. In all cases one sub-pocket completely covers the co-crystallized ligand. Besides the classical overlap-measure of ligand versus predicted active site, we additionally considered the pocket coverage by the co-crystallized ligand. We found that the number of test cases with more than 30% pocket coverage rises from 30% to 74% (PDBbind) and from 28% to 63% (scPDB), respectively, when considering sub-pockets.

In silico prediction of protein binding using molecular dynamic simulations: A future tool for accelerated process development

In silico prediction of protein binding using molecular dynamic simulations: A future tool for accelerated process development

ANCHOR: web server for predicting protein binding regions in disordered proteins

Summary: ANCHOR is a web-based implementation of an original method that takes a single amino acid sequence as an input and predicts protein binding regions that are disordered in isolation but can undergo disorder-to-order transition upon binding. The server incorporates the result of a general disorder prediction method, IUPred and can carry out simple motif searches as well.Availability: The web server is available at http://anchor.enzim.hu. The program package is freely available for academic users.Contact: zsuzsa@enzim.hu

Utilization of estimated physicochemical properties as an integrated part of predicting hepatic clearance in the early drug-discovery stage: Impact of plasma and microsomal binding

Rapid prediction of hepatic clearance for drug candidates plays an important role for decision-making in the early drug-discovery stage. Although knowledge of protein binding in both plasma and microsomal components is needed in the prediction of metabolic clearance from metabolic stability studies, the capacity of protein binding assays are generally lower than those of metabolic stability assays.However, many in silico prediction methods for protein binding are now available and software packages such as ACDLabs, ADMET Predictor and SimCYP incorporate various aspects of in silico predictions relevant to estimating binding and clearance.This has facilitated the use of various estimated or measured physicochemical parameters, relevant to binding, to predict clearance. In this study, prediction of protein binding for 33 drugs was evaluated using various combinations of estimated physicochemical properties. Subsequently, the most accurate estimated protein binding values were used to predict hepatic clearance using the SimCYP software.For the drugs used herein, SimCYP provided the most accurate prediction for protein binding in both plasma and microsomes using physiochemical properties estimated with the ACDLabs software.In conclusion, the use of in silico methods as an integrated part of predicting hepatic clearance in early drug-discovery stage is recommended.

Accurate Calculations of Binding, Folding, and Transfer Free Energies by a Scaled Generalized Born Method

The Poisson-Boltzmann (PB) equation is widely used for modeling solvation effects. The computational cost of PB has restricted its applications largely to single-conformation calculations. The generalized Born (GB) model provides an approximation at substantially reduced cost. Currently the best GB methods reproduce PB results for electrostatic solvation energies with errors at ~5 kcal/mol. When two proteins form a complex, the net electrostatic contributions to the binding free energy are typically of the order of 5 to 10 kcal/mol. Similarly, the net contributions of individual residues to protein folding free energy are < 5 kcal/mol. Clearly in these applications the accuracy of current GB methods is insufficient. Here we present a simple scaling scheme that allows our GB method, GBr6, to reproduce PB results for binding, folding, and transfer free energies with high accuracy. From an ensemble of conformations sampled from molecular dynamics simulations, five were judiciously selected for PB calculations. These PB results were used for scaling GBr6. Tests on the binding free energies of the barnase-barstar, GTPase-WASp, and U1A-U1hpII complexes and on the folding free energy of FKBP show that the effects of point mutations calculated by scaled GBr6 are accurate to within 0.3 kcal/mol of PB results. Similar accuracy was also achieved for the free energies of transfer for ribonuclease Sa and insulin from the crystalline phase to the solution phase at various pH's. This method makes it possible to thoroughly sample the transient-complex ensemble in predicting protein binding rate constants and to incorporate conformational sampling in electrostatic modeling (such as done in the MM-GBSA approach) without loss of accuracy.

Integrating high-content screening and ligand-target prediction to identify mechanism of action

High-content screening is transforming drug discovery by enabling simultaneous measurement of multiple features of cellular phenotype that are relevant to therapeutic and toxic activities of compounds. High-content screening studies typically generate immense datasets of image-based phenotypic information, and how best to mine relevant phenotypic data is an unsolved challenge. Here, we introduce factor analysis as a data-driven tool for defining cell phenotypes and profiling compound activities. This method allows a large data reduction while retaining relevant information, and the data-derived factors used to quantify phenotype have discernable biological meaning. We used factor analysis of cells stained with fluorescent markers of cell cycle state to profile a compound library and cluster the hits into seven phenotypic categories. We then compared phenotypic profiles, chemical similarity and predicted protein binding activities of active compounds. By integrating these different descriptors of measured and potential biological activity, we can effectively draw mechanism-of-action inferences.

Pharmacokinetic and Pharmacodynamic Properties of SB1317, a Potent and Orally Active Multi-Kinase Inhibitor.

Key physicochemical properties such as solubility, lipophilicity (logD7.4, logP) and pKa (the negative log of the acid dissociation constant) can be used to predict protein binding, tissue distribution, and gastrointestinal (GI) absorption. More recent application of computational methods allows even better prediction of compound oral bioavailability from in vitro and/or in silico properties. During the lead optimization process in our multi-kinase inhibitor program a good correlation between physicochemical properties like logD and solubility (determined experimentally at pH 7) versus AUC(0-inf) (Area Under the Curve) was observed for a series of kinase inhibitors. A calculated logD*solubility value between 200 to 400 predicted a reasonably high oral AUC(0-inf) in mouse PharmacoKinetic (PK) studies for this series of compounds, with correlation coefficient of >0.9. The correlation of logD*solubility vs. AUC(0-inf) enabled us to prioritize compounds for PK studies. SB1317, a novel pyrimidine derivative, is an orally active multi-kinase inhibitor that evolved as a lead drug candidate from in vitro and in vivo pharmacokinetic studies. SB1317 is a highly permeable compound, does not undergo active P-glycoprotein transport, and has a solubility of 75 μg/ml. It is metabolically stable in dog and human liver microsomes and unstable in rodent liver microsomes. No P450 inhibition was observed up to 10 μM towards CYP3A4, CYP1A2, CYP2C9 and CYP2C19. In vivo pharmacokinetics in nude mice and Beagle dogs resulted in %F of 12 and 37% respectively. Tumor pharmacokinetics of SB1317 shows increased exposure in tumor than in plasma with a AUC(0-t) tumor/plasma ratio of 3. Excellent oral dose proportionality (10, 20 and 40mg/kg) was observed in both plasma and tumor. SB1317 has exhibited uniformly high plasma protein binding (>99%) across the preclinical species and human. It is the free or unbound portion of the compound in the plasma or at the tissue level that would yield effective anti-tumor activity. (AUC0-inf)unbound/GI50 and Cmax-unbound/GI50 were used as PK/PD surrogates for the measure of SB1317 efficacy. A factor of SB1317 unbound AUC (0-inf)/GI50 above 0.5 results in significant pharmacodynamic (PD) effects in hematological tumor models (MV4-11 and HL-60 engraft model). Linear pharmacokinetic extrapolation from preclinical species to human was determined by allometric scaling (predicted human % F = 36). PK/PD relationships established for SB1317 in preclinical species could form the basis of a PK/PD-driven clinical development program.

Identification of hot regions in protein-protein interactions by sequential pattern mining

BackgroundIdentification of protein interacting sites is an important task in computational molecular biology. As more and more protein sequences are deposited without available structural information, it is strongly desirable to predict protein binding regions by their sequences alone. This paper presents a pattern mining approach to tackle this problem. It is observed that a functional region of protein structures usually consists of several peptide segments linked with large wildcard regions. Thus, the proposed mining technology considers large irregular gaps when growing patterns, in order to find the residues that are simultaneously conserved but largely separated on the sequences. A derived pattern is called a cluster-like pattern since the discovered conserved residues are always grouped into several blocks, which each corresponds to a local conserved region on the protein sequence.ResultsThe experiments conducted in this work demonstrate that the derived long patterns automatically discover the important residues that form one or several hot regions of protein-protein interactions. The methodology is evaluated by conducting experiments on the web server MAGIIC-PRO based on a well known benchmark containing 220 protein chains from 72 distinct complexes. Among the tested 218 proteins, there are 900 sequential blocks discovered, 4.25 blocks per protein chain on average. About 92% of the derived blocks are observed to be clustered in space with at least one of the other blocks, and about 66% of the blocks are found to be near the interface of protein-protein interactions. It is summarized that for about 83% of the tested proteins, at least two interacting blocks can be discovered by this approach.ConclusionThis work aims to demonstrate that the important residues associated with the interface of protein-protein interactions may be automatically discovered by sequential pattern mining. The detected regions possess high conservation and thus are considered as the computational hot regions. This information would be useful to characterizing protein sequences, predicting protein function, finding potential partners, and facilitating protein docking for drug discovery.

QSAR Modeling of Human Serum Protein Binding with Several Modeling Techniques Utilizing Structure−Information Representation

Four modeling techniques, using topological descriptors to represent molecular structure, were employed to produce models of human serum protein binding (% bound) on a data set of 1008 experimental values, carefully screened from publicly available sources. To our knowledge, this data is the largest set on human serum protein binding reported for QSAR modeling. The data was partitioned into a training set of 808 compounds and an external validation test set of 200 compounds. Partitioning was accomplished by clustering the compounds in a structure descriptor space so that random sampling of 20% of the whole data set produced an external test set that is a good representative of the training set with respect to both structure and protein binding values. The four modeling techniques include multiple linear regression (MLR), artificial neural networks (ANN), k-nearest neighbors (kNN), and support vector machines (SVM). With the exception of the MLR model, the ANN, kNN, and SVM QSARs were ensemble models. Training set correlation coefficients and mean absolute error ranged from r2=0.90 and MAE=7.6 for ANN to r2=0.61 and MAE=16.2 for MLR. Prediction results from the validation set yielded correlation coefficients and mean absolute errors which ranged from r2=0.70 and MAE=14.1 for ANN to a low of r2=0.59 and MAE=18.3 for the SVM model. Structure descriptors that contribute significantly to the models are discussed and compared with those found in other published models. For the ANN model, structure descriptor trends with respect to their affects on predicted protein binding can assist the chemist in structure modification during the drug design process.

Protemot: prediction of protein binding sites with automatically extracted geometrical templates

Geometrical analysis of protein tertiary substructures has been an effective approach employed to predict protein binding sites. This article presents the Protemot web server that carries out prediction of protein binding sites based on the structural templates automatically extracted from the crystal structures of protein–ligand complexes in the PDB (Protein Data Bank). The automatic extraction mechanism is essential for creating and maintaining a comprehensive template library that timely accommodates to the new release of PDB as the number of entries continues to grow rapidly. The design of Protemot is also distinctive by the mechanism employed to expedite the analysis process that matches the tertiary substructures on the contour of the query protein with the templates in the library. This expediting mechanism is essential for providing reasonable response time to the user as the number of entries in the template library continues to grow rapidly due to rapid growth of the number of entries in PDB. This article also reports the experiments conducted to evaluate the prediction power delivered by the Protemot web server. Experimental results show that Protemot can deliver a superior prediction power than a web server based on a manually curated template library with insufficient quantity of entries. Availability: .

A Mechanistic Approach for the Scaling of Clearance in Children

Clearance is an important pharmacokinetic concept for scaling dosage, understanding the risks of drug-drug interactions and environmental risk assessment in children. Accurate clearance scaling to children requires prior knowledge of adult clearance mechanisms and the age-dependence of physiological and enzymatic development. The objective of this research was to develop and evaluate ontogeny models that would provide an assessment of the age-dependence of clearance. Using in vitro data and/or in vivo clearance values for children for eight compounds that are eliminated primarily by one process, models for the ontogeny of renal clearance, cytochrome P450 (CYP) 3A4, CYP2E1, CYP1A2, uridine diphosphate glucuronosyltransferase (UGT) 2B7, UGT1A6, sulfonation and biliary clearance were developed. Resulting ontogeny models were evaluated using six compounds that demonstrated elimination via multiple pathways. The proportion of total clearance attributed to each clearance pathway in adults was delineated. Each pathway was individually scaled to the desired age, inclusive of protein-binding prediction, and summed to generate a total plasma clearance for the child under investigation. The paediatric age range included in the study was premature neonates to sub-adults. There was excellent correlation between observed and predicted clearances for the model development (R2 = 0.979) and test sets (Q2 = 0.927). Clearance in premature neonates could also be well predicted (development R2 = 0.951; test Q2 = 0.899). Paediatric clinical trial development could greatly benefit from clearance scaling, particularly in guiding dosing regimens. Furthermore, since the proportion of clearance via different elimination pathways is age-dependent, information could be gained on the developmental extent of drug-drug interactions.

Computational prediction of the plasma protein‐binding percent of diverse pharmaceutical compounds

A nonlinear regression analysis has been applied to develop a new method to predict the plasma protein‐binding percent of structurally diverse pharmaceutical compounds. The analysis included over 300 launched drugs with experimental human plasma protein binding percent data. These drugs were classified according to protonation state and pharmacophore features. The correlation formula for each class is a simple sigmoidal function of variable LogP or LogD. A correlation formula of variable LogD at pH 7.4 with a good correlation coefficient (R‐squared = 0.803) was obtained for neutral and basic drugs, with the exception of zwitterions. A correlation formula using LogP as variable for acidic drugs with one of the specific pharmacophore features gave a good correlation coefficient (R‐squared = 0.786). The method was verified using the protein binding data of 20 compounds that had not been included in the data set to configure the formulas. The correlation coefficient (R‐squared) between the experimental and predicted protein binding percents was 0.830. In conclusion, the method developed and described in this report can provide precise and useful prediction of plasma protein binding percent for new drug candidates. © 2004 Wiley‐Liss, Inc. and the American Pharmacists Association J Pharm Sci 93:1480–1494, 2004