RESUMEN
Deep learning has demonstrated significant potential in advancing state of the art in many problem domains, especially those benefiting from automated feature extraction. Yet, the methodology has seen limited adoption in the field of ligand-based virtual screening (LBVS) as traditional approaches typically require large, target-specific training sets, which limits their value in most prospective applications. Here, we report the development of a neural network architecture and a learning framework designed to yield a generally applicable tool for LBVS. Our approach uses the molecular graph as input and involves learning a representation that places compounds of similar biological profiles in close proximity within a hyperdimensional feature space; this is achieved by simultaneously leveraging historical screening data against a multitude of targets during training. Cosine distance between molecules in this space becomes a general similarity metric and can readily be used to rank order database compounds in LBVS workflows. We demonstrate the resulting model generalizes exceptionally well to compounds and targets not used in its training. In three commonly employed LBVS benchmarks, our method outperforms popular fingerprinting algorithms without the need for any target-specific training. Moreover, we show the learned representation yields superior performance in scaffold hopping tasks and is largely orthogonal to existing fingerprints. Summarily, we have developed and validated a framework for learning a molecular representation that is applicable to LBVS in a target-agnostic fashion, with as few as one query compound. Our approach can also enable organizations to generate additional value from large screening data repositories, and to this end we are making its implementation freely available at https://github.com/totient-bio/gatnn-vs.
Asunto(s)
Algoritmos , Redes Neurales de la Computación , Bases de Datos Factuales , Ligandos , Estudios ProspectivosRESUMEN
DNA-encoded libraries (DELs) are large, pooled collections of compounds in which every library member is attached to a stretch of DNA encoding its complete synthetic history. DEL-based hit discovery involves affinity selection of the library against a protein of interest, whereby compounds retained by the target are subsequently identified by next-generation sequencing of the corresponding DNA tags. When analyzing the resulting data, one typically assumes that sequencing output (i.e., read counts) is proportional to the binding affinity of a given compound, thus enabling hit prioritization and elucidation of any underlying structure-activity relationships (SAR). This assumption, though, tends to be severely confounded by a number of factors, including variable reaction yields, presence of incomplete products masquerading as their intended counterparts, and sequencing noise. In practice, these confounders are often ignored, potentially contributing to low hit validation rates, and universally leading to loss of valuable information. To address this issue, we have developed a method for comprehensively denoising DEL selection outputs. Our method, dubbed "deldenoiser", is based on sparse learning and leverages inputs that are commonly available within a DEL generation and screening workflow. Using simulated and publicly available DEL affinity selection data, we show that "deldenoiser" is not only able to recover and rank true binders much more robustly than read count-based approaches but also that it yields scores, which accurately capture the underlying SAR. The proposed method can, thus, be of significant utility in hit prioritization following DEL screens.
Asunto(s)
ADN/química , Biblioteca de Genes , Aprendizaje AutomáticoRESUMEN
In this study, we present novel insights into the pH-dependent retention behavior of protonated basic solutes in chaotropic chromatography. To this end, two sets of experiments were performed to distinguish between mobile phase pH and ionic strength effects. In the first set, the ionic strength (I) was varied with the concentration of NaPF6 and additives that adjusted the mobile phase pH, while in the second set, I was kept constant by adding the appropriate amount of NaCl. In each set, the retention behavior of 13 analytes was qualitatively examined in 21 chromatographic systems, which were defined by the NaPF6 concentration in their aqueous phases (1-50mM) and the pH of their mobile phases (2, 3 or 4); the acetonitrile content was fixed at 40%. The addition of NaCl significantly reduced the differences among retention factors at studied pH values due to the effect of the Na+ ions on PF6-adsorption to the stationary phase and the magnitude of the consequential development of the surface potential. A quantitative description of the observed phenomenon was obtained by an extended thermodynamic approach. The contribution of ion-pair formation in the stationary phase to the retention of the solutes was confirmed across models at the studied pH values in the set with varying I. In the systems with a constant I, the shielding effect of the Na+ ions on the surface charge lowered the attractive surface potential and diminished the aforementioned interactions and hence the effect of the mobile phase pH on analyte retention. Eventually, we developed a readily interpretable empirical retention model that simultaneously takes into account analyte molecular structures and the most relevant chromatographic factors. Its coefficients have clear physical meaning, and owing to its good predictive capabilities, the model could be successfully used to clarify the contributions of analyte molecular structures and chromatographic factors to the specific processes underlying separation in chaotropic chromatography.
Asunto(s)
Cromatografía Líquida de Alta Presión/métodos , Fosfatos/química , Acetonitrilos/química , Adsorción , Concentración de Iones de Hidrógeno , Iones/química , Concentración Osmolar , Cloruro de Sodio/química , TermodinámicaRESUMEN
In this article, retention modeling of eight aminopyridines (synthesized and characterized at the Faculty of Pharmacy) in reversed-phase high performance liquid chromatography (RP-HPLC) was performed. No data related to their retention in the RP-HPLC system were found. Knowing that, it was recognized as very important to describe their retention behavior. The influences of pH of the mobile phase and the organic modifier content on the retention factors were investigated. Two theoretical models for the dependence of retention factor of organic modifier content were tested. Then, the most reliable and accurate prediction of log k was created, testing multiple linear regression model-quantitative structure-retention relationships (MLR-QSRR) and support vector regression machine-quantitative structure-retention relationships (SVM-QSRR). Initially, 400 descriptors were calculated, but four of them (POM, log D, M-SZX/RZX and m-RPCG) were included in the models. SVM-QSRR performed significantly better than the MLR model. Apart from aminopyridines, four structurally similar substances (indapamide, gliclazide, sulfamethoxazole and furosemide) were followed in the same chromatographic system. They were used as external validation set for the QSRR model (it performed well within its applicability domain, which was defined using a bounding box approach). After having described retention of eight aminopyridines with both theoretical and QSRR models, further investigations in this field can be conducted.
Asunto(s)
Aminopiridinas/aislamiento & purificación , Cromatografía Líquida de Alta Presión/métodos , Cromatografía de Fase Inversa/métodos , Modelos Estadísticos , Agua , Acetonitrilos , Aminopiridinas/síntesis química , Cromatografía Líquida de Alta Presión/estadística & datos numéricos , Cromatografía de Fase Inversa/estadística & datos numéricos , Concentración de Iones de Hidrógeno , Soluciones , SolventesRESUMEN
The aim of this study was to systematically investigate the phenomena affecting the retention behavior of structurally diverse basic drugs in ion-interaction chromatographic systems with chaotropic additives. To this end, the influence of three factors was studied: pH value of the aqueous phase, concentration of sodium hexafluorophosphate, and content of acetonitrile in the mobile phase. Mobile phase pH was found to affect the thermodynamic equilibria in the studied system beyond its effects on the analytes' ionization state. Specifically, increasing pH from 2 to 4 led to longer retention times, even with analytes which remain completely protonated. An explanation for this phenomenon was sought by studying the adsorption behavior of acetonitrile and chaotropic additive onto stationary phase. It was shown that the magnitude of the developed surface potential, which significantly affects retention - increases with pH, and that this can be attributed to the larger surface excess of acetonitrile. To study how analytes' structural properties influence their retention, quantitative structure-retention modeling was performed next. A support vector machine regression model was developed, relating mobile phase constituents and structural descriptors with retention data. While the ETA_EtaP_B_RC and XlogP can be considered as molecular descriptors which describe factors affecting retention in any RP-HPLC system, TDB9p and RDF45p are molecular descriptors which account for spatial arrangement of polarizable atoms and they can clearly relate to analytes' behavior on the stationary phase surface, where the electrostatic potential develops. Complementarity of analytes' structure with that of the electric double layer can be seen as a key factor influencing their retention behavior. Structural diversity of analytes and good predictive capabilities over a range of experimental conditions make the established model a useful tool in predicting retention behavior in the studied chromatographic system.
Asunto(s)
Técnicas de Química Analítica/métodos , Cromatografía Líquida de Alta Presión , Iones/química , Acetonitrilos/química , Adsorción , Concentración de Iones de Hidrógeno , Indicadores y Reactivos , Fosfatos/química , Electricidad Estática , Termodinámica , Agua/químicaRESUMEN
The aim of this study was to examine the interaction of the chaotropic salts of different position in Hofmeister series (CF3COONa, NaClO4, NaPF6) added to the mobile phase with the stationary phases of different hydrophobicity (C8 and C18 XTerra(®) columns), as well as their common influence on the retention behavior of pramipexole and its structurally related impurities. The extended thermodynamic approach enabled the understanding of the underlying separation mechanism. Comparing six different column-salt systems it was observed that general system hydrophobicity presented by salt chaotropicity and column hydrophobicity favors stationary phase ion-pairing over the ion-pair formation in the eluent. Further, an attempt was made to describe the influence of analytes' nature on their retention behavior in such chromatographic systems. An analysis is performed in order to select and elucidate the molecular descriptors (electrostatical, quantum-chemical, geometrical, topological, and constitutional) that best explain the experimental evidence and findings obtained by the thermodynamic approach. The results of this analysis suggest that analytes' charge distribution and its complementarity to the structure of the electric double layer formed on the surface of the stationary phase upon the addition of chaotropic additives can be useful for understanding the differences in retention of structurally related analytes. These findings provide a novel understanding of the interactions between all the components of the chromatographic system containing chaotropic additive and a good basis for further investigations suggesting the development of generally applicable predictors in structure-retention relationship studies in related chromatographic systems.
Asunto(s)
Benzotiazoles/análisis , Cromatografía Liquida , Sales (Química)/química , Benzotiazoles/aislamiento & purificación , Fluoroacetatos/química , Interacciones Hidrofóbicas e Hidrofílicas , Percloratos/química , Pramipexol , Teoría Cuántica , Compuestos de Sodio/química , Electricidad Estática , TermodinámicaRESUMEN
Enhancer of Zeste Homolog 2 (EZH2) is a SET domain protein lysine methyltransferase (PKMT) which has recently emerged as a chemically tractable and therapeutically promising epigenetic target, evidenced by the discovery and characterization of potent and highly selective EZH2 inhibitors. However, no experimental structures of the inhibitors co-crystallized to EZH2 have been resolved, and the structural basis for their activity and selectivity remains unknown. Considering the need to minimize cross-reactivity between prospective PKMT inhibitors, much can be learned from understanding the molecular basis for selective inhibition of EZH2. Thus, to elucidate the binding of small-molecule inhibitors to EZH2, we have developed a model of its fully-formed cofactor binding site and used it to carry out molecular dynamics simulations of protein-ligand complexes, followed by molecular mechanics/generalized born surface area calculations. The obtained results are in good agreement with biochemical inhibition data and reflect the structure-activity relationships of known ligands. Our findings suggest that the variable and flexible post-SET domain plays an important role in inhibitor binding, allowing possibly distinct binding modes of inhibitors with only small variations in their structure. Insights from this study present a good basis for design of novel and optimization of existing compounds targeting the cofactor binding site of EZH2.
Asunto(s)
Epigénesis Genética , Histonas/química , Complejo Represivo Polycomb 2/química , Relación Estructura-Actividad , Secuencia de Aminoácidos , Sitios de Unión , Proteína Potenciadora del Homólogo Zeste 2 , Humanos , Simulación de Dinámica Molecular , Complejo Represivo Polycomb 2/antagonistas & inhibidores , Bibliotecas de Moléculas PequeñasRESUMEN
In this work, we present a novel approach for the development of models for prediction of aqueous solubility, based on the implementation of an algorithm for the automatic adjustment of descriptor's relative importance (AARI) in counter-propagation artificial neural networks (CPANN). Using this approach, the interpretability of the models based on artificial neural networks, which are traditionally considered as "black box" models, was significantly improved. For the development of the model, a data set consisting of 374 diverse drug-like molecules, divided into training (n=280) and test (n=94) sets using self-organizing maps, was used. Heuristic method was applied in preselecting a small number of the most significant descriptors to serve as inputs for CPANN training. The performances of the final model based on 7 descriptors for prediction of solubility were satisfactory for both training (RMSEP(train)=0.668) and test set (RMSEP(test)=0.679). The model was found to be a highly interpretable in terms of solubility, as well as rationalizing structural features that could have an impact on the solubility of the compounds investigated. Therefore, the proposed approach can significantly enhance model usability by giving guidance for structural modifications of compounds with the aim of improving solubility in the early phase of drug discovery.