A BERT-based pretraining model for extracting molecular structural information from a SMILES sequence.

Among the various molecular properties and their combinations, it is a costly process to obtain the desired molecular properties through theory or experiment. Using machine learning to analyze molecular structure features and to predict molecular properties is a potentially efficient alternative for accelerating the prediction of molecular properties. In this study, we analyze molecular properties through the molecular structure from the perspective of machine learning. We use SMILES sequences as inputs to an artificial neural network in extracting molecular structural features and predicting molecular properties. A SMILES sequence comprises symbols representing molecular structures. To address the problem that a SMILES sequence is different from actual molecular structural data, we propose a pretraining model for a SMILES sequence based on the BERT model, which is widely used in natural language processing, such that the model learns to extract the molecular structural information contained in the SMILES sequence. In an experiment, we first pretrain the proposed model with 100,000 SMILES sequences and then use the pretrained model to predict molecular properties on 22 data sets and the odor characteristics of molecules (98 types of odor descriptor). The experimental results show that our proposed pretraining model effectively improves the performance of molecular property prediction SCIENTIFIC CONTRIBUTION: The 2-encoder pretraining is proposed by focusing on the lower dependency of symbols to the contextual environment in a SMILES than one in a natural language sentence and the corresponding of one compound to multiple SMILES sequences. The model pretrained with 2-encoder shows higher robustness in tasks of molecular properties prediction compared to BERT which is adept at natural language.

Investigation of the structure-odor relationship using a Transformer model.

The relationships between molecular structures and their properties are subtle and complex, and the properties of odor are no exception. Molecules with similar structures, such as a molecule and its optical isomer, may have completely different odors, whereas molecules with completely distinct structures may have similar odors. Many works have attempted to explain the molecular structure-odor relationship from chemical and data-driven perspectives. The Transformer model is widely used in natural language processing and computer vision, and the attention mechanism included in the Transformer model can identify relationships between inputs and outputs. In this paper, we describe the construction of a Transformer model for predicting molecular properties and interpreting the prediction results. The SMILES data of 100,000 molecules are collected and used to predict the existence of molecular substructures, and our proposed model achieves an F1 value of 0.98. The attention matrix is visualized to investigate the substructure annotation performance of the attention mechanism, and we find that certain atoms in the target substructures are accurately annotated. Finally, we collect 4462 molecules and their odor descriptors and use the proposed model to infer 98 odor descriptors, obtaining an average F1 value of 0.33. For the 19 odor descriptors that achieved F1 values greater than 0.45, we also attempt to summarize the relationship between the molecular substructures and odor quality through the attention matrix.

Machine-Learning-Based Olfactometer: Prediction of Odor Perception from Physicochemical Features of Odorant Molecules.

Gas chromatography/olfactometry (GC/O) has been used in various fields as a valuable method to identify odor-active components from a complex mixture. Since human assessors are employed as detectors to obtain the olfactory perception of separated odorants, the GC/O technique is limited by its subjectivity, variability, and high cost of the trained panelists. Here, we present a proof-of-concept model by which odor information can be obtained by machine-learning-based prediction from molecular parameters (MPs) of odorant molecules. The odor prediction models were established using a database of flavors and fragrances including 1026 odorants and corresponding verbal odor descriptors (ODs). Physicochemical parameters of the odorant molecules were acquired by use of molecular calculation software (DRAGON). Ten representative ODs were selected to build the prediction models based on their high frequency of occurrence in the database. The features of the MPs were extracted via either unsupervised (principal component analysis) or supervised (Boruta, BR) approaches and then used as input to calibrate machine-learning models. Predictions were performed by various machine-learning approaches such as support vector machine (SVM), random forest, and extreme learning machine. All models were optimized via parameter tuning and their prediction accuracies were compared. A SVM model combined with feature extraction by BR-C (confirmed only) was found to afford the best results with an accuracy of 97.08%. Validation of the models was verified by using the GC/O data of an apple sample for comparison between the predicted and measured results. The prediction models can be used as an auxiliary tool in the existing GC/O by suggesting possible OD candidates to the panelists and thus helping to give more objective and correct judgment. In addition, a machine-based GC/O in which the panelist is no longer needed might be expected after further development of the proposed odor prediction technique.