Combining QSAR Modeling and Text-Mining Techniques to Link Chemical Structures and Carcinogenic Modes of Action

Deep neural networks can directly learn from chemical structures without extensive, user-driven selection of descriptors in order to predict molecular properties/activities with high reliability. But these approaches typically require large training sets to learn the endpoint-specific structural features and ensure reasonable prediction accuracy. Even though large datasets are becoming the new normal in drug discovery, especially when it comes to high-throughput screening or metabolomics datasets, one should also consider smaller datasets with challenging endpoints to model and forecast. Thus, it would be highly relevant to better utilize the tremendous compendium of unlabeled compounds from publicly-available datasets for improving the model performances for the user’s particular series of compounds. In this study, we propose the Molecular Prediction Model Fine-Tuning (MolPMoFiT) approach, an effective transfer learning method based on self-supervised pre-training + task-specific fine-tuning for QSPR/QSAR modeling. A large-scale molecular structure prediction model is pre-trained using one million unlabeled molecules from ChEMBL in a self-supervised learning manner, and can then be fine-tuned on various QSPR/QSAR tasks for smaller chemical datasets with specific endpoints. Herein, the method is evaluated on four benchmark datasets (lipophilicity, FreeSolv, HIV, and blood-brain barrier penetration). The results showed the method can achieve strong performances for all four datasets compared to other state-of-the-art machine learning modeling techniques reported in the literature so far.

Download Full-text

IDO as a drug target for cancer immunotherapy: recent developments in IDO inhibitors discovery

RSC Advances ◽

10.1039/c5ra25046c ◽

2016 ◽

Vol 6 (9) ◽

pp. 7575-7581 ◽

Cited By ~ 29

Author(s):

Shan Qian ◽

Man Zhang ◽

Quanlong Chen ◽

Yanying He ◽

Wei Wang ◽

...

Keyword(s):

Cancer Immunotherapy ◽

Small Molecule ◽

Drug Target ◽

Immune Escape ◽

Chemical Structures ◽

Modes Of Action ◽

Recent Advances ◽

Recent Developments ◽

Ido Inhibitors

This review highlights the recent advances in research related to the role of IDO in immune escape in cancer and novel small-molecule IDO inhibitors with an emphasis on their chemical structures and modes of action.

Download Full-text

Hydrophobic Property of (R)-3 Amidinophenylalanine Inhibitors Contributes to their Inhibition Constants with Thrombin Enzyme

Science and Technology Development Journal ◽

10.32508/stdj.v22i3.1684 ◽

2019 ◽

Vol 22 (3) ◽

Author(s):

Hien Van Nguyen ◽

Van Thi Bich Pham ◽

Hao Minh Hoang

Keyword(s):

Goodness Of Fit ◽

Blood Clot ◽

Thrombin Inhibitors ◽

Coagulation Cascade ◽

Hydrophobic Property ◽

Qsar Modeling ◽

Data Set ◽

Chemical Structures ◽

2D Qsar ◽

Inhibition Constants

Introduction: Thrombin is the key enzyme of fibrin formation in the blood coagulation cascade. Thrombin is released by the hydrolysis of prothrombinase which is generated from factor Xa and factor Va in the presence of calcium ion and phospholipid. The inhibition of thrombin is of therapeutic interest in blood clot treatment. Currently, potent thrombin inhibitors of (R)-3- amidinophenylalanine, derived from benzamidine-containing amino acid, have been developed so far. In order to quantitatively express a relationship between chemical structures and inhibition constants (Ki with thrombin enzyme in a data set of (R)-3-amidinophenylalanine inhibitors), we developed a quantitative structure-activity relationship (QSAR) modeling from a group of 60 (R)-3- amidinophenylalanine inhibitors. Methods: A database containing chemical structures of 60 inhibitors and their Ki values was put into molecular operating environment (MOE) 2008.10 software, and the two-dimensional (2D) physicochemical descriptors were numerically calculated. After removing the irrelevant descriptors, a QSAR modeling was developed from the 2D-descriptors and Ki values by using the partial least squares (PLS) regression method. Results: The results showed that the hydrophobic property, reflected through n-octanol/water partition coefficient (P) of a drug molecule, contributes mainly to Ki values with thrombin.The statistic parameters that give the information about the goodness of fit of a 2D-QSAR model (such as squared correlation coefficient of R2 = 0.791, root mean square error (RMSE) = 0.443, cross-validated Q2 cv = 0.762, and cross-validated RMSEcv = 0.473) were statistically obtained for a training set (60 inhibitors). The R2 and RMSE values were obtained by using a developed model for the testing set (9 inhibitors) ; the total set has statistically significant parameters. Furthermore, the 2D-QSAR modeling was also applied to predict the Ki values of the 69 inhibitors. A linear relationship was found between the experimental and predicted pKi values of the inhibitors. Conclusion: The results support the promising application of established 2D-QSAR modeling in the prediction and design of new (R)-3-amidinophenylalanine candidates in the pharmaceutical industry.

Download Full-text

Inductive Transfer Learning for Molecular Activity Prediction: Next-Gen QSAR Models with MolPMoFiT

10.26434/chemrxiv.9978743 ◽

2020 ◽

Cited By ~ 1

Author(s):

Xinhao Li ◽

Denis Fourches

Keyword(s):

Transfer Learning ◽

High Throughput Screening ◽

Structure Prediction ◽

Large Scale ◽

High Reliability ◽

Structural Features ◽

Fine Tuning ◽

Qsar Modeling ◽

Chemical Structures ◽

Effective Transfer

Deep neural networks can directly learn from chemical structures without extensive, user-driven selection of descriptors in order to predict molecular properties/activities with high reliability. But these approaches typically require large training sets to learn the endpoint-specific structural features and ensure reasonable prediction accuracy. Even though large datasets are becoming the new normal in drug discovery, especially when it comes to high-throughput screening or metabolomics datasets, one should also consider smaller datasets with challenging endpoints to model and forecast. Thus, it would be highly relevant to better utilize the tremendous compendium of unlabeled compounds from publicly-available datasets for improving the model performances for the user’s particular series of compounds. In this study, we propose the Molecular Prediction Model Fine-Tuning (MolPMoFiT) approach, an effective transfer learning method based on self-supervised pre-training + task-specific fine-tuning for QSPR/QSAR modeling. A large-scale molecular structure prediction model is pre-trained using one million unlabeled molecules from ChEMBL in a self-supervised learning manner, and can then be fine-tuned on various QSPR/QSAR tasks for smaller chemical datasets with specific endpoints. Herein, the method is evaluated on four benchmark datasets (lipophilicity, FreeSolv, HIV, and blood-brain barrier penetration). The results showed the method can achieve strong performances for all four datasets compared to other state-of-the-art machine learning modeling techniques reported in the literature so far.

Download Full-text

An Analysis of QSAR Research Based on Machine Learning Concepts

Current Drug Discovery Technologies ◽

10.2174/1570163817666200316104404 ◽

2020 ◽

Vol 17 ◽

Cited By ~ 2

Author(s):

Mohammad Reza Keyvanpour ◽

Mehrnoush Barani Shirzad

Keyword(s):

Machine Learning ◽

Learning Strategies ◽

Biological Activities ◽

Analytical Framework ◽

Quantitative Structure Activity Relationship ◽

Machine Learning Algorithms ◽

Machine Learning Techniques ◽

Future Research ◽

Qsar Modeling ◽

Chemical Structures

: Quantitative Structure–Activity Relationship (QSAR) is a popular approach developed to correlate chemical molecules with their biological activities based on their chemical structures. Machine learning techniques have proved to be promising solutions to QSAR modeling. Due to the significant role of machine learning strategies in QSAR modeling, this area of research has attracted much attention from researchers. A considerable amount of literature has been published on machine learning based QSAR modeling methodologies whilst this domain still suffers from lack of a recent and comprehensive analysis of these algorithms. This study systematically reviews the application of machine learning algorithms in QSAR, aiming to provide an analytical framework. For this purpose, we present a framework called ‘ML-QSAR‘. This framework has been designed for future research to: a)facilitate the selection of proper strategies among existing algorithms according to the application area requirements, b) help to develop and ameliorate current methods and c) providing a platform to study existing methodologies comparatively. In ML-QSAR, first a structured categorization is depicted which studied the QSAR modeling research based on machine models. Then several criteria are introduced in order to assess the models. Finally, inspired by aforementioned criteria the qualitative analysis is carried out.

Download Full-text

Computation of Vertex-Based Topological Descriptors of Organometallic Monolayers of TM 3 C 12 S 12

Journal of Mathematics ◽

10.1155/2021/8572049 ◽

2021 ◽

Vol 2021 ◽

pp. 1-7

Author(s):

Dalal Alrowaili ◽

Faraha Ashraf ◽

Rifaqat Ali ◽

Arsalan Shoukat ◽

Aqila Shaheen ◽

...

Keyword(s):

Quantitative Structure Activity Relationship ◽

Topological Descriptors ◽

Quantitative Structure ◽

Structure Property ◽

Biological Structure ◽

Qsar Modeling ◽

Chemical Structures ◽

Structure Activity ◽

Structure Property Relationship ◽

Mathematical Relationships

Topological descriptors are mathematical values related to chemical structures which are associated with different physicochemical properties. The use of topological descriptors has a great contribution in the field of quantitative structure-property relationship (QSPR) and quantitative structure-activity relationship (QSAR) modeling. These are mathematical relationships between different molecular properties or biological activity and some other physicochemical or structural properties. In this article, we calculate few vertex degree-based topological indices/descriptors of the organometallic monolayer structure. At present, the numerical programming of the biological structure with topological descriptors is increasing in consequence in invigorating science, bioinformatics, and pharmaceutics.

Download Full-text

Chemical structure of diamond-like carbon thin films

Proceedings, annual meeting, Electron Microscopy Society of America ◽

10.1017/s042482010017668x ◽

1990 ◽

Vol 48 (4) ◽

pp. 710-711

Author(s):

N.-H. Cho ◽

K.M. Krishnan ◽

D.B. Bogy

Keyword(s):

Electrical Resistivity ◽

Chemical Structure ◽

Diamond Like Carbon ◽

Chemical Structures ◽

Sputtering Power ◽

Data Aquisition ◽

Equilibrium Phases ◽

Electron Energy Loss ◽

Power Densities ◽

Preparation Techniques

Diamond-like carbon (DLC) films have attracted much attention due to their useful properties and applications. These properties are quite variable depending on film preparation techniques and conditions, DLC is a metastable state formed from highly non-equilibrium phases during the condensation of ionized particles. The nature of the films is therefore strongly dependent on their particular chemical structures. In this study, electron energy loss spectroscopy (EELS) was used to investigate how the chemical bonding configurations of DLC films vary as a function of sputtering power densities. The electrical resistivity of the films was determined, and related to their chemical structure.DLC films with a thickness of about 300Å were prepared at 0.1, 1.1, 2.1, and 10.0 watts/cm2, respectively, on NaCl substrates by d.c. magnetron sputtering. EEL spectra were obtained from diamond, graphite, and the films using a JEOL 200 CX electron microscope operating at 200 kV. A Gatan parallel EEL spectrometer and a Kevex data aquisition system were used to analyze the energy distribution of transmitted electrons. The electrical resistivity of the films was measured by the four point probe method.

Download Full-text