Accurate Physical Property Predictions via Deep Learning

Neural networks and deep learning have been successfully applied to tackle problems in drug discovery with increasing accuracy over time. There are still many challenges and opportunities to improve molecular property predictions with satisfactory accuracy even further. Here, we proposed a deep-learning architecture model, namely Bidirectional long short-term memory with Channel and Spatial Attention network (BCSA), of which the training process is fully data-driven and end to end. It is based on data augmentation and SMILES tokenization technology without relying on auxiliary knowledge, such as complex spatial structure. In addition, our model takes the advantages of the long- and short-term memory network (LSTM) in sequence processing. The embedded channel and spatial attention modules in turn specifically identify the prime factors in the SMILES sequence for predicting properties. The model was further improved by Bayesian optimization. In this work, we demonstrate that the trained BSCA model is capable of predicting aqueous solubility. Furthermore, our proposed method shows noticeable superiorities and competitiveness in predicting oil–water partition coefficient, when compared with state-of-the-art graphs models, including graph convoluted network (GCN), message-passing neural network (MPNN), and AttentiveFP.     

三嗪类化合物溶解度参数及毒性构－效关系

测定了１２种三嗪类化合物的水溶解度，辛醇水分配系数和对发光菌的毒性，并用分子连结性指数建立了预测三嗪类化合物的溶解度，辛醇水分配系数及对发光菌毒性的定量结构活性相关方程，其中１０种化合物文献中未见报道。     

三嗪类化合物溶解度参数及毒性构—效关系

测定了１２种三嗪类化合物的水溶解度，辛醇水分配系数和对发光菌的毒性，并用分子连结性指数建立了预测三嗪类化合物的溶解度，辛醇水分配系数及对发光菌毒性的定量结构活性相关方程，其中１０种化合物献中未见报道。e     

Prediction of pH-dependent aqueous solubility of Histone Deacetylase (HDAC) inhibitors1

Recently we developed a model for prediction of pH-dependent aqueous solubility of drugs and drug like molecules. In the present work, the model was applied on a series of novel Histone Deacetylases (HDAC) inhibitors discovered at TopoTarget. The applicability of our model was evaluated on the series of HDAC inhibitors by use of Self-Organizing Maps (SOM) and 2D-projection of the HDAC inhibitors on the chemical space of the training data set of the artificial neural network (ANN) module. The model was refined for the particular chemical space of interest, which led to two modifications in the training data set of the ANN. The performance of the original and the two modified versions of the model were evaluated against the commercial software from Simulations-plus and pH-dependent solubility measurements for representative compounds of the series. The results of the evaluation indicate that one can develop models that are more accurate in predicting differences in the solubility of structurally very similar compounds than models that have been trained on structurally unbiased, diverse data sets. Such ‘tailor-made’ models have the potential to become trustworthy enough to replace time-consuming and expensive medium- and high-throughput solubility experiments by providing results of similar or even better quality.     

Using artificial neural networks for prediction of organic acid partition coefficients

Aqueous/organic phase partition coefficients of organic acids were predicted using an artificial neural network (ANN) algorithm taking benzoic acid derivatives as examples. The partition coefficients were determined by extraction of the acids from aqueous salt solutions with hydrophilic solvents (BuⁿOH, BuⁱOH, and Bu^tOH). Using the ANN approach makes it possible to obtain quantitative information on the values of the title parameters. Published in Russian in Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 2, pp. 207—212, February, 2006.     

液液萃取中半挥发性有机物正辛醇/水分配系数与回收率相关性分析

该文讨论了水中目标有机物的回收率与正辛醇/水分配系数之间的联系,结果表明半挥发性有机污染物在水中的回收率与正辛醇/水分配系数（lgKow）间存在一定的规律：当lgKow＜5时,回收率为16.1%~56.1%,当lgKow＞5时,回收率为59.5%~124.9%;用溶解度（lgS）与89种有机污染物的回收率进行分析,得出与正辛醇/水分配系数（lgKow）一致的结论：当lgS＜-0.60时,回收率为59.5%~124.9%,当lgS＞-0.6时,回收率为16.1%~56.1%;采用实际水样的实验室空白加标结果对所得结论进行验证,其结果显示与上述结论一致,即当污染物的lgKow＞5,或者lgS＜-0.6时,其液液萃取的回收率为85.3%~115.9%。     

Prediction of Chromatography Conditions for Purification in Organic Synthesis Using Deep Learning

In this research, a process for developing normal-phase liquid chromatography solvent systems has been proposed. In contrast to the development of conditions via thin-layer chromatography (TLC), this process is based on the architecture of two hierarchically connected neural network-based components. Using a large database of reaction procedures allows those two components to perform an essential role in the machine-learning-based prediction of chromatographic purification conditions, i.e., solvents and the ratio between solvents. In our paper, we build two datasets and test various molecular vectorization approaches, such as extended-connectivity fingerprints, learned embedding, and auto-encoders along with different types of deep neural networks to demonstrate a novel method for modeling chromatographic solvent systems employing two neural networks in sequence. Afterward, we present our findings and provide insights on the most effective methods for solving prediction tasks. Our approach results in a system of two neural networks with long short-term memory (LSTM)-based auto-encoders, where the first predicts solvent labels (by reaching the classification accuracy of 0.950 ± 0.001) and in the case of two solvents, the second one predicts the ratio between two solvents (R² metric equal to 0.982 ± 0.001). Our approach can be used as a guidance instrument in laboratories to accelerate scouting for suitable chromatography conditions.     

Aqueous Solubilities, Infinite Dilution Activity Coefficients and Octanol–Water Partition Coefficients of Tricyclic Analogs of Acyclovir

Solubilities of tricyclic analogs of acyclovir have been determined in water at 25, 35, and 45°C and in octanol, water-saturated octanol, and octanol-saturated water at 25°C. Octanol-water partition coefficients were determined at 25°C. Melting temperatures and molar enthalpies of fusion were measured. Activity coefficients in water, octanol, and in aqueous octanol solutions were determined and are discussed. The effect of hydrophilic and hydrophobic substituents in the tricyclic analogs on their thermodynamic properties are discussed. The standard Gibbs energy of transfer between the saturated phases were found to correlate with known values of the melting point of the solvents and the solubilities of the solute. For a number of the compounds examined, correlations between the minimum inhibitory concentration against the herpes simplex virus type 1 (HSV-1) and type 2 (HSV-2), varicella-zoster virus (VZV), thymidine kinase-deficient (TK^–) strains of VZV and were established. Detailed conclusions have been derived concerning the relationships between the structure and the thermodynamic parameters of the compounds examined.     

几类醚化物的水溶性、碱水溶解性和酸解活性

本文对几类典型醚类化合物的水溶性、碱水(5%的Na₂SiO₃·9H₂O)溶解性和酸解活性进行了初步研究,查阅了这些醚类化合物的水溶性,并测定了其碱水溶解性、临界碱水不可溶醚当量以及在微量酸存在条件下的酸解活性,研究结果表明,对一般醚类化合物而言,当醚键中不含p-π共轭时其临界水不溶醚当量为116,临界碱水不溶醚当量为102;当醚键中含p-π共轭时其临界水不溶醚当量在56-100之间,临界碱水不溶醚当量在56-72之间,这些数据对光/热成像用活性醚化物阻溶/促溶剂的分子设计以及碱显影成像制版具有重大的指导意义。     

New data on the solubility of amorphous silica in organic compound-water solutions and a new model for the thermodynamic behavior of aqueous silica in aqueous complex solutions

Experimental solubilities of amorphous silica in several aqueous electrolyte solutions and in aqueous solutions of organic compounds, and theoretical considerations concerning cavity formation, electrostriction collapse, ion solvation, and long- and short-range interaction of the solvated ions with one another⁽¹⁾ permit the calculation of the partial excess free energies and the activity coefficients of aqueous silica. It is shown that, in the case of non-dissociated aqueous organic solutions, the variation of log m (SiO₂) with the reciprocal of the dielectric constant of the solution is described by a single linear equation independent of the nature of the organic compound. For aqueous electrolyte solutions, a specific linear relationship between log m (SiO₂) and the reciprocal of the dielectric constant occurs for each electrolyte. The success of the equation in reproducing the experimental solubilities of amorphous silica in aqueous solutions of electrolytes and organic compounds supports previous evidence indicating a polar charge distribution in the solvated SiO₂ molecule. Our data permit the calculation of the effective local charge of dissolved SiO₂ molecules and of the short-range interaction parameters between SiO₂ and various ions. The proposed equation of state can be used to calculate the affinity of reactions among SiO₂ minerals and complex aqueous solutions.     

Novel topological index for research on structure-property relationships of complex organic compounds

Based on bonding parameters such as Yang's Electronegative Force Gauge Y(i), electronic number of valence layer Z(i), number of combined hydrogen atoms h(i), number of bonding electron b(i), and quantum number such as the highest main quantum number of valence layer n(i), a novel atomic valence delta(i) (Y) is defined and a novel topological index (1)chi(Y) is derived from the atomic valence. The atomic valence is defined as delta(i) (Y) = (Z(i) - h(i))b(i)/n(i) (2)Y(i), while the topological index is expressed as (1)chi(Y) summation operator (i,j=1) (m) (delta(i) (Y)delta(j) (Y))(-1/2). Subsequently, the index (1)chi(Y) is utilized to study the structure-property relationships of complex organic compounds. The results of correlativity showed that the index is highly and extensively correlated with such properties as solubility of phenyl chlorides, gas chromatographic retention index of alkoxyl silanes, and toxicity of heterocyclic nitrogen-containing compounds. Moreover, predicted values are quite consistent with experimental ones when the index is employed to predict the partition coefficient (log P) of fatty alcohols, phenyl chlorides, and barbitals. Compared to the topological indices reported in the literature, the universality and reliability of (1)chi(Y) to the properties of complex organic compounds have been distinctively improved, and its calculating process is simple and convenient.     

Effect of nucleic acid contamination on partitioning of proteins in two-phase PEG-Dextran system

Downstream processing of bioproducts results in considerable losses of compounds of interest in a large number of cases. For the intracellular enzyme tartrate dehydrogenase, an analysis of the laboratory process for enzyme recovery revealed that maximum losses occur in the initial stages of purification when the enzyme is separated from nucleic acids and other undesirable enzymes. Hence, aqueous twophase extraction was studied to investigate the separation of several enzymes from nucleic acids. Single-component and binary equilibria for three commercially available enzymes (bovine serum albumin, trypsin, chymotrypsin) and yeast RNA were studied in a two-phase system consisting of dextran and polyethylene glycol (PEG). The effects of pH and concentrations of the components and salts (NaC1) were investigated.