Third International School-Seminar "From Empirical to Predictive chemistry" Abstracts

October 12, 2018

PANDAS AND SKLEARN FOR QSAR MODELING

Ramil Nugmanov

Quantitative structure-activity relationship approach is the use of mathematical models to quantitatively describe the relationship between molecular structure, represented by some characteristics called descriptors, and biological activity (QSAR) or any other property (QSPR). Machine learning methods are used for establishing mathematical dependency between property and descriptors. Special techniques like cross-validation are used to assess models’ predictive ability, control overfi...

Read More

March 14, 2018

DATABASE AND QSPR ANALYSIS OF ANIONIC DYES FOR TEXTILE COLORATION

Felix Telegin

Present study is aiming at development of database of anionic dyes for textile coloration and QSPR analysis by the use of fragment approach and software NASAWIN (Baskin et al., 2002). Series of papers of well-known research groups selected in current research make a valuable fundamental contribution to systematic synthesis and experimental study of different properties of dyeings on natural cotton, wool, silk and synthetic polyamide. Database for acid azo dyes is collected from the research:...

Read More

March 14, 2018

BASICS OF PYTHON PROGRAMMING

Ilseyar Alimova

The programming language Python is widely used in the field of data science. The syntax simplicity of this language makes it possible to improve developer productivity and readability of code. In addition, the advantages of Python are: cross-platforming, dynamic typing, the availability of a large number of libraries, support for both object oriented and procedural programming, combination with other languages, distribution under an open license. In the field of Chemoinformatics, Python is al...

Read More

March 14, 2018

CONFORMATIONAL REARRANGEMENT OF H-BONDING {S(O)-CH2-CH2-OH∙∙∙}2 SYNTHON: A QUANTUM-TOPOLOGICAL ANALYSIS

Olga Lodochnikova

For the triclinic and monoclinic modifications of diastereomeric pinanyl sulfoxides co-crystal a conformational restructuring of a stable hydrogen-bonded synthon from “unfolded” to “folded” form by transition from the temperature 293 К to 150 К was ascertained. The driving force of this restructuring is the tendency to realization of S=O…S=O interactions, which show up in the low-temperature phases of both polymorphs and are well-reproducible using the methods of quantum chemistry (DFT, B97-D...

Read More

March 13, 2018

Introduction to RDKit

Pavel Polishchuk

RDKit chemoinformatics toolkit is one of the most powerful and functon-rich open-source (BSD license) tools to process and manipulate small molecules.

In this introductory course the following topics will be covered: 1. Reading/writing molecules
2. Working with molecules. Looping over bonds and atoms, get/set properties, magic properties.
3. 3D structures and conformers
4. Substructure search
5. Descriptors calculation
6. Structural and pharmacophore fingerprints
7. Simi...

Read More

March 12, 2018

CGRtools – library for CGR based reactions management and analysis

Ramil Nugmanov

The Condensed Graphs of Reaction (CGR) is effective approach to handle structural information on reactions or other chemical transformations (e.g., equilibria) [1], as well as pseudo-transformations (like Matched Molecular Pairs) [2]. CGR represents chemical transformation as a single graph that contains both conventional chemical bonds (e.g. single, double, etc.) and so called “dynamic” bonds and atoms characterizing chemical transformations, i.e. breakage or formation of bond or changing ...

Read More

March 10, 2018

QSPR ANALYSIS OF BODIPY DYES IN CHEMISTRY AND HIGH-TECH ENGINEERING

Felix Telegin

The research is aiming at structure-property analysis of different families of BODIPY dyes created elsewhere or collected from different systematic research available in the literature. Method of QSPR analysis is based on fragment approach (Baskin et al, 2002) taking into account sub-structural fragments of atoms in a form of chains, branches and cycles of different size. Final results are described by multiple linear regression models and correspondent statistics. The results of the researc...

Read More

March 10, 2018

Application of Deep Learning Neural Networks in Chemoinformatics: Advantages and Prospects

Igor Baskin

This lecture is devoted to the advances in the application of neural networks with deep architecture in chemoinformatics.

Read More

March 10, 2018

Chemography analysis of drug-like chemical space

Alexandre Varnek

Here, we discuss "universal' Generative Topographic Maps of the ChEMBL database. They can efficiently be used to predict pharmacological profiles of biologically-active compounds and to extract some "privilage" structural motifs for the ligands against particular target.

Read More

March 10, 2018

Condensed Graph of Reaction – an efficient approach for reactions mining

Timur Madzhidov

In chemoinformatics, a chemical reaction represents a difficult object because it involves several molecular graphs of two types (reactants and products) and its kinetic and thermodynamic parameters depend on experimental conditions (catalyst, solvent, temperature, etc). This complexity often prevents applying to chemical reactions most of approaches developed for individual molecules. The complexity can be significantly reduced in the framework of the Condensed Graph of Reaction (CGR) approa...

Read More