Discovery of Small-Molecule Autophagy Inhibitors by Disrupting the Protein-Protein Interactions Involving Autophagy-Related 5.

One possible strategy for modulating autophagy is to disrupt the critical protein-protein interactions (PPIs) formed during this process. Our attention is on the autophagy-related 12 (ATG12)-autophagy-related 5 (ATG5)-autophagy-related 16-like 1 (ATG16L1) heterotrimer complex, which is responsible for ATG8 translocation from ATG3 to phosphatidylethanolamine. In this work, we discovered a compound with an (E)-3-(2-furanylmethylene)-2-pyrrolidinone core moiety (T1742) that blocked the ATG5-ATG16L1 and ATG5-TECAIR interactions in the in vitro binding assay (IC50 = 1-2 µM) and also exhibited autophagy inhibition in cellular assays. The possible binding mode of T1742 to ATG5 was predicted through molecular modeling, and a batch of derivatives sharing essentially the same core moiety were synthesized and tested. The outcomes of the in vitro binding assay and the flow cytometry assay of those newly synthesized compounds were generally consistent. This work has validated our central hypothesis that small-molecule inhibitors of the PPIs involving ATG5 can tune down autophagy effectively, and their pharmaceutical potential may be further explored.

Public Data Set of Protein-Ligand Dissociation Kinetic Constants for Quantitative Structure-Kinetics Relationship Studies.

Protein-ligand binding affinity reflects the equilibrium thermodynamics of the protein-ligand binding process. Binding/unbinding kinetics is the other side of the coin. Computational models for interpreting the quantitative structure-kinetics relationship (QSKR) aim at predicting protein-ligand binding/unbinding kinetics based on protein structure, ligand structure, or their complex structure, which in principle can provide a more rational basis for structure-based drug design. Thus far, most of the public data sets used for deriving such QSKR models are rather limited in sample size and structural diversity. To tackle this problem, we have compiled a set of 680 protein-ligand complexes with experimental dissociation rate constants (k off), which were mainly curated from the references accumulated for updating our PDBbind database. Three-dimensional structure of each protein-ligand complex in this data set was either retrieved from the Protein Data Bank or carefully modeled based on a proper template. The entire data set covers 155 types of protein, with their dissociation kinetic constants (k off) spanning nearly 10 orders of magnitude. To the best of our knowledge, this data set is the largest of its kind reported publicly. Utilizing this data set, we derived a random forest (RF) model based on protein-ligand atom pair descriptors for predicting k off values. We also demonstrated that utilizing modeled structures as additional training samples will benefit the model performance. The RF model with mixed structures can serve as a baseline for testifying other more sophisticated QSKR models. The whole data set, namely, PDBbind-koff-2020, is available for free download at our PDBbind-CN web site (http://www.pdbbind.org.cn/download.php).

Tapping on the Black Box: How Is the Scoring Power of a Machine-Learning Scoring Function Dependent on the Training Set?

In recent years, protein-ligand interaction scoring functions derived through machine-learning are repeatedly reported to outperform conventional scoring functions. However, several published studies have questioned that the superior performance of machine-learning scoring functions is dependent on the overlap between the training set and the test set. In order to examine the true power of machine-learning algorithms in scoring function formulation, we have conducted a systematic study of six off-the-shelf machine-learning algorithms, including Bayesian Ridge Regression (BRR), Decision Tree (DT), K-Nearest Neighbors (KNN), Multilayer Perceptron (MLP), Linear Support Vector Regression (L-SVR), and Random Forest (RF). Model scoring functions were derived with these machine-learning algorithms on various training sets selected from over 3700 protein-ligand complexes in the PDBbind refined set (version 2016). All resulting scoring functions were then applied to the CASF-2016 test set to validate their scoring power. In our first series of trial, the size of the training set was fixed; while the overall similarity between the training set and the test set was varied systematically. In our second series of trial, the overall similarity between the training set and the test set was fixed, while the size of the training set was varied. Our results indicate that the performance of those machine-learning models are more or less dependent on the contents or the size of the training set, where the RF model demonstrates the best learning capability. In contrast, the performance of three conventional scoring functions (i.e., ChemScore, ASP, and X-Score) is basically insensitive to the use of different training sets. Therefore, one has to consider not only "hard overlap" but also "soft overlap" between the training set and the test set in order to evaluate machine-learning scoring functions. In this spirit, we have complied data sets based on the PDBbind refined set by removing redundant samples under several similarity thresholds. Scoring functions developers are encouraged to employ them as standard training sets if they want to evaluate their new models on the CASF-2016 benchmark.

Revisiting the Relationship Between Correlation Coefficient, Confidence Level, and Sample Size.

In the field of computational chemistry, it is a very common task to compare the predictive power of theoretical models with Pearson correlation coefficients. A general understanding is that larger sample sizes lead to increased precision. However, what is the minimum sample size required for comparing two models? This issue has not been well addressed in this field. To the best of our knowledge, the only serious study of this kind was published by Carlson in 2013 [ J. Chem. Inf. Model. 2013 , 53 1837 - 1841 ], where they proposed a method for estimating the minimum sample size required by this task. Considering how a benchmark comparison is conducted in reality, we want to point out that (i) the possible intercorrelation between two models should not be neglected and (ii) the one-sided test is more reasonable because comparison direction is known a priori. Carlson's method has significantly overestimated the required minimum sample size due to these two issues. Here, we will describe a more appropriate method based on Dunn and Clark's test statistic, and we have designed an extensive numerical test to validate our method. The minimum sample sizes required by comparing two models under various conditions are computed with our method. Our study has shown that the required minimum sample size is determined by several factors, including confidence, power, correlation coefficients as well as the intercorrelation between two models. As a rule of thumb, a couple of hundred samples are sufficient at 90% confidence or above for comparing two models producing meaningful R values.

Comparative Assessment of Scoring Functions: The CASF-2016 Update.

In structure-based drug design, scoring functions are often employed to evaluate protein-ligand interactions. A variety of scoring functions have been developed so far, and thus, some objective benchmarks are desired for assessing their strength and weakness. The comparative assessment of scoring functions (CASF) benchmark developed by us provides an answer to this demand. CASF is designed as a "scoring benchmark", where the scoring process is decoupled from the docking process to depict the performance of scoring function more precisely. Here, we describe the latest update of this benchmark, i.e., CASF-2016. Each scoring function is still evaluated by four metrics, including "scoring power", "ranking power", "docking power", and "screening power". Nevertheless, the evaluation methods have been improved considerably in several aspects. A new test set is compiled, which consists of 285 protein-ligand complexes with high-quality crystal structures and reliable binding constants. A panel of 25 scoring functions are tested on CASF-2016 as a demonstration. Our results reveal that the performance of current scoring functions is more promising in terms of docking power than scoring, ranking, and screening power. Scoring power is somewhat correlated with ranking power, so are docking power and screening power. The results obtained on CASF-2016 may provide valuable guidance for the end users to make smart choices among available scoring functions. Moreover, CASF is created as an open-access benchmark so that other researchers can utilize it to test a wider range of scoring functions. The complete CASF-2016 benchmark will be released on the PDBbind-CN web server ( http://www.pdbbind-cn.org/casf.asp/ ) once this article is published.

Assessing protein-ligand interaction scoring functions with the CASF-2013 benchmark.

Scoring functions are a group of computational methods widely applied in structure-based drug design for fast evaluation of protein-ligand interactions. To date, a whole spectrum of scoring functions have been developed based on different assumptions or algorithms. Therefore, it is important to both the end users and the developers of scoring functions that their performance be objectively assessed. We have developed the comparative assessment of scoring functions (CASF) benchmark as an open-access solution for scoring function evaluation. The latest CASF-2013 benchmark enables evaluation of the so-called 'scoring power', 'ranking power', 'docking power', and 'screening power' of a given scoring function with a high-quality test set of 195 complexes formed between diverse protein molecules and their small-molecule ligands. Evaluation results of the standard scoring functions implemented in several mainstream software programs (including Schrödinger, MOE, Discovery Studio, SYBYL, and GOLD) are provided as reference. This benchmark has become popular among the scoring function community since its first release. In this protocol, we provide detailed descriptions of the data files included in the CASF-2013 package and step-by-step instructions on how to conduct the performance tests with the ready-to-use computer scripts included in the package. This protocol is expected to lower the technical hurdles in front of new and existing users of the CASF-2013 benchmark. On a standard desktop workstation, it takes roughly half an hour to complete the whole evaluation procedure for one scoring function, once the required inputs, i.e., the results computed on the test set, are ready to use.

Enhance the performance of current scoring functions with the aid of 3D protein-ligand interaction fingerprints.

BACKGROUND: In structure-based drug design, binding affinity prediction remains as a challenging goal for current scoring functions. Development of target-biased scoring functions provides a new possibility for tackling this problem, but this approach is also associated with certain technical difficulties. We previously reported the Knowledge-Guided Scoring (KGS) method as an alternative approach (BMC Bioinformatics, 2010, 11, 193-208). The key idea is to compute the binding affinity of a given protein-ligand complex based on the known binding data of an appropriate reference complex, so the error in binding affinity prediction can be reduced effectively. RESULTS: In this study, we have developed an upgraded version, i.e. KGS2, by employing 3D protein-ligand interaction fingerprints in reference selection. KGS2 was evaluated in combination with four scoring functions (X-Score, ChemPLP, ASP, and GoldScore) on five drug targets (HIV-1 protease, carbonic anhydrase 2, beta-secretase 1, beta-trypsin, and checkpoint kinase 1). In the in situ scoring test, considerable improvements were observed in most cases after application of KGS2. Besides, the performance of KGS2 was always better than KGS in all cases. In the more challenging molecular docking test, application of KGS2 also led to improved structure-activity relationship in some cases. CONCLUSIONS: KGS2 can be applied as a convenient "add-on" to current scoring functions without the need to re-engineer them, and its application is not limited to certain target proteins as customized scoring functions. As an interpolation method, its accuracy in principle can be improved further with the increasing knowledge of protein-ligand complex structures and binding affinity data. We expect that KGS2 will become a practical tool for enhancing the performance of current scoring functions in binding affinity prediction. The KGS2 software is available upon contacting the authors.

Forging the Basis for Developing Protein-Ligand Interaction Scoring Functions.

In structure-based drug design, scoring functions are widely used for fast evaluation of protein-ligand interactions. They are often applied in combination with molecular docking and de novo design methods. Since the early 1990s, a whole spectrum of protein-ligand interaction scoring functions have been developed. Regardless of their technical difference, scoring functions all need data sets combining protein-ligand complex structures and binding affinity data for parametrization and validation. However, data sets of this kind used to be rather limited in terms of size and quality. On the other hand, standard metrics for evaluating scoring function used to be ambiguous. Scoring functions are often tested in molecular docking or even virtual screening trials, which do not directly reflect the genuine quality of scoring functions. Collectively, these underlying obstacles have impeded the invention of more advanced scoring functions. In this Account, we describe our long-lasting efforts to overcome these obstacles, which involve two related projects. On the first project, we have created the PDBbind database. It is the first database that systematically annotates the protein-ligand complexes in the Protein Data Bank (PDB) with experimental binding data. This database has been updated annually since its first public release in 2004. The latest release (version 2016) provides binding data for 16 179 biomolecular complexes in PDB. Data sets provided by PDBbind have been applied to many computational and statistical studies on protein-ligand interaction and various subjects. In particular, it has become a major data resource for scoring function development. On the second project, we have established the Comparative Assessment of Scoring Functions (CASF) benchmark for scoring function evaluation. Our key idea is to decouple the "scoring" process from the "sampling" process, so scoring functions can be tested in a relatively pure context to reflect their quality. In our latest work on this track, i.e. CASF-2013, the performance of a scoring function was quantified in four aspects, including "scoring power", "ranking power", "docking power", and "screening power". All four performance tests were conducted on a test set containing 195 high-quality protein-ligand complexes selected from PDBbind. A panel of 20 standard scoring functions were tested as demonstration. Importantly, CASF is designed to be an open-access benchmark, with which scoring functions developed by different researchers can be compared on the same grounds. Indeed, it has become a popular choice for scoring function validation in recent years. Despite the considerable progress that has been made so far, the performance of today's scoring functions still does not meet people's expectations in many aspects. There is a constant demand for more advanced scoring functions. Our efforts have helped to overcome some obstacles underlying scoring function development so that the researchers in this field can move forward faster. We will continue to improve the PDBbind database and the CASF benchmark in the future to keep them as useful community resources.

AutoT&T v.2: An Efficient and Versatile Tool for Lead Structure Generation and Optimization.

UNLABELLED: In structure-based drug design, automated de novo design methods are helpful tools for lead discovery as well as lead optimization. In a previous study ( J. Chem. Inf. MODEL: 2011 , 51 , 1474 - 1491 ) we reported a new de novo design method, namely, Automatic Tailoring and Transplanting (AutoT&T). It overcomes some intrinsic problems in conventional fragment-based buildup methods. In this study, we describe an upgraded version, namely, AutoT&T2. Structural operations conducted by AutoT&T2 have been largely optimized by introducing several new algorithms. As a result, its overall speed in multiround optimization jobs has been improved by a few thousand fold. With this improvement, it is now practical to conduct structural crossover among multiple lead molecules using AutoT&T2. Three different test cases are described in this study that demonstrate the new features and versatile applications of AutoT&T2. The AutoT&T2 software suite is available to the public. Besides, a Web portal for running AutoT&T2 online is provided at http://www.sioc-ccbg.ac.cn/software/att2 for testing.

Alloscore: a method for predicting allosteric ligand-protein interactions.

UNLABELLED: Allosteric ligands have increasingly gained attention as potential therapeutic agents due to their higher target selectivity and lower toxicity compared with classic orthosteric ligands. Despite the great interest in the development of allosteric drugs as a new tactic in drug discovery, the understanding of the ligand-protein interactions underlying allosteric binding represents a key challenge. Herein, we introduce Alloscore, a web server that predicts the binding affinities of allosteric ligand-protein interactions. This method exhibits prominent performance in describing allosteric binding and could be useful in allosteric virtual screening and the structural optimization of allosteric agonists/antagonists. AVAILABILITY AND IMPLEMENTATION: The Alloscore server and tutorials are freely available at http://mdl.shsmu.edu.cn/alloscore CONTACT: jian.zhang@sjtu.edu.cn SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.