Assima Rakhimbekova
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Assima Rakhimbekova has a total of 41 co-authors. Here are the top co-authors by number of shared publications:
- Timur I. Madzhidov
Chemistry Solutions, Life Science Solutions, Elsevier, Nielsen House, John Smith Drive, Oxford OX4 2WB, United Kingdom
5 shared publications - Alexandre Varnek
Laboratory of Chemoinformatics, UMR 7140 CNRS, University of Strasbourg, 4, rue B. Pascal, Strasbourg, France, 67000
3 shared publications - Timur Rustemovich Gimadiev
A.M. Butlerov Institute of Chemistry, Kazan Federal University, Kazan, Russia
3 shared publications - Igor I. Baskin
Faculty of Physics, M.V. Lomonosov Moscow State University, Moscow, Russia
2 shared publications - Ramil I. Nugmanov
Laboratory of Chemoinformatics and Molecular Modeling, Butlerov Institute of Chemistry, Kazan Federal University, 18, Kremlyovskaya str., 420008 Kazan, Russia
2 shared publications - Tagir Akhmetshin
University of Strasbourg
2 shared publications
Latest Publications
- Open Access
Conditional Variational AutoEncoder to Predict Suitable Conditions for Hydrogenation Reactions
Reaction conditions (RCs) are a crucial part of reaction definition, and their accurate prediction is an important component of chemical synthesis planning. The existence of multiple combinations of RCs capable of achieving the desired result complicates the task of condition recommendation. Herein, we propose a conditional variational autoencoder (CVAE) generative model to predict suitable RCs. The CVAE model has been customized to generate diverse sets of valid conditions, ensuring high flexibility and accuracy, while circumventing the necessity for enumeration or combinatorial search of potential RCs. The efficacy of the CVAE approaches was evaluated using hydrogenation reactions and other H2-mediated reactions, predicting the set of catalysts, additives (acid, base, and catalytic poison), ranges of temperature, and pressure. The CVAE models predicted conditions with different “heads”, each corresponding to specific condition components, and their respective losses. CVAE models were tested on two datasets: a small one containing 31K reactions with 2232 potential conditions’ combinations and a big one having 196K reactions with ~7 × 1042 potential conditions’ combinations to evaluate the model’s ability to predict varying complexity and diversity conditions. To optimize the accuracy of the models, we experimented with three latent distribution variants—Gaussian (g-CVAE), Riemannian Normalizing Flow (rnf-CVAE), and Hyperspherical Uniform (h-CVAE). In our experiments, the h-CVAE model demonstrated robust overall performance, making it the optimal choice for scenarios requiring high accuracy across multiple top-k predictions. Benchmarking analyses demonstrated the high performance of the CVAE models compared to state-of-the-art reaction condition prediction approaches.
Open AccessConditional Variational AutoEncoder to Predict Suitable Conditions for Hydrogenation Reactions
Reaction conditions (RCs) are a crucial part of reaction definition, and their accurate prediction is an important component of chemical synthesis planning. The existence of multiple combinations of RCs capable of achieving the desired result complicates the task of condition recommendation. Herein, we propose a conditional variational autoencoder (CVAE) generative model to predict suitable RCs. The CVAE model has been customized to generate diverse sets of valid conditions, ensuring high flexibility and accuracy, while circumventing the necessity for enumeration or combinatorial search of potential RCs. The efficacy of the CVAE approaches was evaluated using hydrogenation reactions and other H2-mediated reactions, predicting the set of catalysts, additives (acid, base, and catalytic poison), ranges of temperature, and pressure. The CVAE models predicted conditions with different “heads”, each corresponding to specific condition components, and their respective losses. CVAE models were tested on two datasets: a small one containing 31K reactions with 2232 potential conditions’ combinations and a big one having 196K reactions with ~7 × 1042 potential conditions’ combinations to evaluate the model’s ability to predict varying complexity and diversity conditions. To optimize the accuracy of the models, we experimented with three latent distribution variants—Gaussian (g-CVAE), Riemannian Normalizing Flow (rnf-CVAE), and Hyperspherical Uniform (h-CVAE). In our experiments, the h-CVAE model demonstrated robust overall performance, making it the optimal choice for scenarios requiring high accuracy across multiple top-k predictions. Benchmarking analyses demonstrated the high performance of the CVAE models compared to state-of-the-art reaction condition prediction approaches.
- Open Access
- 8 May 2022
- Published by Publishing House ABV Press in Tumors of female reproductive system
- Vol. 18 (1) , 76-86
A new algorithm to assess the risk of malignancy in premenopausal patients with pelvic mass
Objective: to elaborate a new algorithm, based on serum CA125, HE4 and age, to assess the risk of malignancy in premenopausal patients with pelvic mass, which performs better than Risk of Ovarian Malignancy Algorithm (ROMA).Materials and methods. The training dataset included 284 premenopausal patients operated because of the presence of pelvic mass, out of which there were 249 patients with benign diseases and 35 patients with malignant or borderline tumors. A novel algorithm, based on serum HE4, CA125 and patient’s age as variables, has been developed. This algorithm was named Risk of Ovarian Cancer Kazan Index (ROCK-I). The validating dataset consisted of 227 consecutively operated premenopausal patients with pelvic mass out of which there were 193 cases of benign diseases, 27 cancers and 7 borderline ovarian tumors (BOT).Results. In the validating dataset ROCK-I and ROMA demonstrated 15 and 30 false positive results respectively. Thus the specificities of ROCK-I and ROMA were 92.2 % and 84.5 %, respectively (р = 0.017). The sensitivities of ROCK-I and ROMA for the joint group of Epithelial ovarian cancers (EOC) (all stages) together with BOT stage IC2–III were 96.3 % and 92.6 %, respectively (p = 0.55). For all malignant disease (all stages) together with BOT stage IC2–III the sensitivities were 90 % and 86.7 %, respectively (p = 0.69). The positive predictive values of ROCK-I and ROMA were 65.1 % and 47.4 %, respectively (p = 0.07). When the scenario of discrimination “benign disease vs the joint group of EOC (all stages) together with BOT stage IC2–III” was used, ROC-AUC of ROCK-I, ROMA and CA125 were 0.988, 0.946 and 0.937. The difference in ROC-AUC between ROCK-I and CA125 was statistically significant (p = 0.01) while the difference between ROMA and CA125 was not (p = 0.79).Conclusion. The proposed ROCK-I has demonstrated greater diagnostic performance than both ROMA and CA125 in the analyzed dataset. If an independent validation shows similar or even slightly lower superiority of ROCK-I over ROMA, it may provide a new basis of routine-use of HE4 in premenopausal patients with pelvic mass.
- 8 May 2022
- Published by Publishing House ABV Press in Tumors of female reproductive system
- Vol. 18 (1) , 76-86
Open AccessA new algorithm to assess the risk of malignancy in premenopausal patients with pelvic mass
Objective: to elaborate a new algorithm, based on serum CA125, HE4 and age, to assess the risk of malignancy in premenopausal patients with pelvic mass, which performs better than Risk of Ovarian Malignancy Algorithm (ROMA).Materials and methods. The training dataset included 284 premenopausal patients operated because of the presence of pelvic mass, out of which there were 249 patients with benign diseases and 35 patients with malignant or borderline tumors. A novel algorithm, based on serum HE4, CA125 and patient’s age as variables, has been developed. This algorithm was named Risk of Ovarian Cancer Kazan Index (ROCK-I). The validating dataset consisted of 227 consecutively operated premenopausal patients with pelvic mass out of which there were 193 cases of benign diseases, 27 cancers and 7 borderline ovarian tumors (BOT).Results. In the validating dataset ROCK-I and ROMA demonstrated 15 and 30 false positive results respectively. Thus the specificities of ROCK-I and ROMA were 92.2 % and 84.5 %, respectively (р = 0.017). The sensitivities of ROCK-I and ROMA for the joint group of Epithelial ovarian cancers (EOC) (all stages) together with BOT stage IC2–III were 96.3 % and 92.6 %, respectively (p = 0.55). For all malignant disease (all stages) together with BOT stage IC2–III the sensitivities were 90 % and 86.7 %, respectively (p = 0.69). The positive predictive values of ROCK-I and ROMA were 65.1 % and 47.4 %, respectively (p = 0.07). When the scenario of discrimination “benign disease vs the joint group of EOC (all stages) together with BOT stage IC2–III” was used, ROC-AUC of ROCK-I, ROMA and CA125 were 0.988, 0.946 and 0.937. The difference in ROC-AUC between ROCK-I and CA125 was statistically significant (p = 0.01) while the difference between ROMA and CA125 was not (p = 0.79).Conclusion. The proposed ROCK-I has demonstrated greater diagnostic performance than both ROMA and CA125 in the analyzed dataset. If an independent validation shows similar or even slightly lower superiority of ROCK-I over ROMA, it may provide a new basis of routine-use of HE4 in premenopausal patients with pelvic mass.
- Open Access
Efficient Design of Peptide-Binding Polymers Using Active Learning Approaches
Active learning (AL) has become a subject of active recent research both in industry and academia as an efficient approach for rapid design and discovery of novel chemicals, materials, and polymers. The key advantages of this approach relate to its ability to (i) employ relatively small datasets for model development, (ii) iterate between model development and model assessment using small external datasets that can be either generated in focused experimental studies or formed from subsets of the initial training data, and (iii) progressively evolve models toward increasingly more reliable predictions and the identification of novel chemicals with the desired properties. Herein, we first compared various AL protocols for their effectiveness in finding biologically active molecules using synthetic datasets. We have investigated the dependency of AL performance on the size of the initial training set, the relative complexity of the task, and the choice of the initial training dataset. We found that AL techniques as applied to regression modeling offer no benefits over random search, while AL used for classification tasks performs better than models built for randomly selected training sets but still quite far from perfect. Using the best performing AL protocol, we have assessed the applicability of AL for the discovery of polymeric micelle formulations for poorly soluble drugs. Finally, the best performing AL approach was employed to discover and experimentally validate novel binding polymers for a case study of asialoglycoprotein receptor (ASGPR).
Open AccessEfficient Design of Peptide-Binding Polymers Using Active Learning Approaches
Active learning (AL) has become a subject of active recent research both in industry and academia as an efficient approach for rapid design and discovery of novel chemicals, materials, and polymers. The key advantages of this approach relate to its ability to (i) employ relatively small datasets for model development, (ii) iterate between model development and model assessment using small external datasets that can be either generated in focused experimental studies or formed from subsets of the initial training data, and (iii) progressively evolve models toward increasingly more reliable predictions and the identification of novel chemicals with the desired properties. Herein, we first compared various AL protocols for their effectiveness in finding biologically active molecules using synthetic datasets. We have investigated the dependency of AL performance on the size of the initial training set, the relative complexity of the task, and the choice of the initial training dataset. We found that AL techniques as applied to regression modeling offer no benefits over random search, while AL used for classification tasks performs better than models built for randomly selected training sets but still quite far from perfect. Using the best performing AL protocol, we have assessed the applicability of AL for the discovery of polymeric micelle formulations for poorly soluble drugs. Finally, the best performing AL approach was employed to discover and experimentally validate novel binding polymers for a case study of asialoglycoprotein receptor (ASGPR).
- 19 February 2021
- Published by Taylor & Francis in SAR and QSAR in Environmental Research
- Vol. 32 (3) , 207-219
Cross-validation strategies in QSPR modelling of chemical reactions
In this article, we consider cross-validation of the quantitative structure-property relationship models for reactions and show that the conventional k-fold cross-validation (CV) procedure gives an ‘optimistically’ biased assessment of prediction performance. To address this issue, we suggest two strategies of model cross-validation, ‘transformation-out’ CV, and ‘solvent-out’ CV. Unlike the conventional k-fold cross-validation approach that does not consider the nature of objects, the proposed procedures provide an unbiased estimation of the predictive performance of the models for novel types of structural transformations in chemical reactions and reactions going under new conditions. Both the suggested strategies have been applied to predict the rate constants of bimolecular elimination and nucleophilic substitution reactions, and Diels-Alder cycloaddition. All suggested cross-validation methodologies and tutorial are implemented in the open-source software package CIMtools (https://github.com/cimm-kzn/CIMtools).
- 19 February 2021
- Published by Taylor & Francis in SAR and QSAR in Environmental Research
- Vol. 32 (3) , 207-219
Cross-validation strategies in QSPR modelling of chemical reactions
In this article, we consider cross-validation of the quantitative structure-property relationship models for reactions and show that the conventional k-fold cross-validation (CV) procedure gives an ‘optimistically’ biased assessment of prediction performance. To address this issue, we suggest two strategies of model cross-validation, ‘transformation-out’ CV, and ‘solvent-out’ CV. Unlike the conventional k-fold cross-validation approach that does not consider the nature of objects, the proposed procedures provide an unbiased estimation of the predictive performance of the models for novel types of structural transformations in chemical reactions and reactions going under new conditions. Both the suggested strategies have been applied to predict the rate constants of bimolecular elimination and nucleophilic substitution reactions, and Diels-Alder cycloaddition. All suggested cross-validation methodologies and tutorial are implemented in the open-source software package CIMtools (https://github.com/cimm-kzn/CIMtools).
- Open Access
- 30 September 2020
- Published by American Chemical Society (ACS) in ChemRxiv
Atom-to-Atom Mapping: A Benchmarking Study of Popular Mapping Algorithms and Consensus Strategies
Here, we discuss a reaction standardization protocol followed by a comparison of popular Atom-to-atom mapping (AAM) tools (ChemAxon, Indigo, RDTool, NextMove and RXNMapper) as well as some consensus AAM strategies. For this purpose, a dataset of 1851 manually curated and mapped reactions was prepared (the Golden dataset) and used as a reference set. It has been found that RXNMapper possesses the highest accuracy, despite the fact that it has some clear disadvantages. Finally, RXNMapper was selected as the best tool, and it was applied to map the USPTO dataset. The standardization protocol used to prepare the data, as well as the data itself are available in the GitHub repository https://github.com/Laboratoire-de-Chemoinformatique.
- 30 September 2020
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessAtom-to-Atom Mapping: A Benchmarking Study of Popular Mapping Algorithms and Consensus Strategies
Here, we discuss a reaction standardization protocol followed by a comparison of popular Atom-to-atom mapping (AAM) tools (ChemAxon, Indigo, RDTool, NextMove and RXNMapper) as well as some consensus AAM strategies. For this purpose, a dataset of 1851 manually curated and mapped reactions was prepared (the Golden dataset) and used as a reference set. It has been found that RXNMapper possesses the highest accuracy, despite the fact that it has some clear disadvantages. Finally, RXNMapper was selected as the best tool, and it was applied to map the USPTO dataset. The standardization protocol used to prepare the data, as well as the data itself are available in the GitHub repository https://github.com/Laboratoire-de-Chemoinformatique.
- 22 October 2019
- Published by American Chemical Society (ACS) in Journal of Chemical Information and Modeling
- Vol. 59 (11) , 4569-4576
Conjugated Quantitative Structure–Property Relationship Models: Application to Simultaneous Prediction of Tautomeric Equilibrium Constants and Acidity of Molecules
Here, we describe a concept of conjugated models for several properties (activities) linked by a strict mathematical relationship. This relationship can be directly integrated analytically into the Ridge Regression (RR) algorithm or accounted for in a special case of “twin” neural networks (NN). Developed approaches were applied to the modelling of the logarithm of the prototropic tautomeric constant (logKT) which can be expressed as the difference between the acidity constants (pKa) of two related tautomers. Both conjugated and individual RR and NN models for logKT and pKa were developed. The modelling set included 639 tautomeric constants and 2371 acidity constants of organic molecules in various solvents. Descriptors vector for each reaction resulted from the concatenation of structural descriptors and some parameters for reaction conditions. As a former, atom-centered substructural fragments describing acid sites in tautomer molecules were used. The latter were automatically identified using the Condensed Graph of Reaction approach. Conjugated models performed similarly to the best individual models for logKT and pKa. At the same time, the physically grounded relationship between logKT and pKa was respected only for conjugated but not individual models.
- 22 October 2019
- Published by American Chemical Society (ACS) in Journal of Chemical Information and Modeling
- Vol. 59 (11) , 4569-4576
Conjugated Quantitative Structure–Property Relationship Models: Application to Simultaneous Prediction of Tautomeric Equilibrium Constants and Acidity of Molecules
Here, we describe a concept of conjugated models for several properties (activities) linked by a strict mathematical relationship. This relationship can be directly integrated analytically into the Ridge Regression (RR) algorithm or accounted for in a special case of “twin” neural networks (NN). Developed approaches were applied to the modelling of the logarithm of the prototropic tautomeric constant (logKT) which can be expressed as the difference between the acidity constants (pKa) of two related tautomers. Both conjugated and individual RR and NN models for logKT and pKa were developed. The modelling set included 639 tautomeric constants and 2371 acidity constants of organic molecules in various solvents. Descriptors vector for each reaction resulted from the concatenation of structural descriptors and some parameters for reaction conditions. As a former, atom-centered substructural fragments describing acid sites in tautomer molecules were used. The latter were automatically identified using the Condensed Graph of Reaction approach. Conjugated models performed similarly to the best individual models for logKT and pKa. At the same time, the physically grounded relationship between logKT and pKa was respected only for conjugated but not individual models.
- 17 October 2017
- Published by Wiley in Journal of Raman Spectroscopy
- Vol. 49 (2) , 207-214
SERS investigation of neocuproine adsorption on silver: Influence of electrode potential on methyl groups
Study of molecules adsorption on charged surfaces is important for biologically relevant substances where the potential at the interface such as living cell membrane is a significant parameter in the processes of their transportation or transmembrane penetration. In this work, a hybrid optical/electrochemical surface‐enhanced Raman scattering (SERS) technique was applied to get new insight into the adsorption state and conformational equilibrium of neocuproine, which serves as a nucleic acid biosensor in clinical diagnostics and has biological activity towards several types of carcinoma. The density functional theory calculations performed for several rotational conformations and their anion radicals were used to determine the geometrical and energetic characteristics, to evaluate the rotational barrier, to obtain the vibrational assignment, and to consider the metal‐adsorbate charge transfer. The dependence of SERS spectra on surface potential is ascribed to a change of the rotational dynamics of methyl groups from hindered to almost free at potentials ≤−200 mV. It is demonstrated for the first time that SERS spectroscopy is capable to recognize the surface species, which differ in the methyl group internal rotation.
- 17 October 2017
- Published by Wiley in Journal of Raman Spectroscopy
- Vol. 49 (2) , 207-214
SERS investigation of neocuproine adsorption on silver: Influence of electrode potential on methyl groups
Study of molecules adsorption on charged surfaces is important for biologically relevant substances where the potential at the interface such as living cell membrane is a significant parameter in the processes of their transportation or transmembrane penetration. In this work, a hybrid optical/electrochemical surface‐enhanced Raman scattering (SERS) technique was applied to get new insight into the adsorption state and conformational equilibrium of neocuproine, which serves as a nucleic acid biosensor in clinical diagnostics and has biological activity towards several types of carcinoma. The density functional theory calculations performed for several rotational conformations and their anion radicals were used to determine the geometrical and energetic characteristics, to evaluate the rotational barrier, to obtain the vibrational assignment, and to consider the metal‐adsorbate charge transfer. The dependence of SERS spectra on surface potential is ascribed to a change of the rotational dynamics of methyl groups from hindered to almost free at potentials ≤−200 mV. It is demonstrated for the first time that SERS spectroscopy is capable to recognize the surface species, which differ in the methyl group internal rotation.
- Open Access
- 4 April 2016
- Published by Royal Society of Chemistry (RSC) in Organic & Biomolecular Chemistry
- Vol. 14 (18) , 4233-4238
Highly selective binding of methyl orange dye by cationic water-soluble pillar[5]arenes
A new water-soluble pillar[5]arene with an amide fragment and triethylammonium groups was synthesized by our original method of aminolysis of the ester groups. Using UV-spectroscopy, it is shown that cationic pillar[5]arenes are able to selectively form 1 : 1 complexes with some hydrophobic anions: the guests with bulky uncharged or negatively charged substituents hindering entry into the macrocycle cavity. Highly selective binding of the most lipophilic guest, methyl orange dye, in the form of organic anion salts by positively charged water-soluble pillar[5]arenes was detected. In the case of the azo dye the appropriate Kass values were 10–100-fold higher than those calculated for the other sulfonic acid derivatives studied. The 2D NMR NOESY 1H–1H spectroscopy confirms the formation of the inclusion complex: negative charge sulfonate head is outside the cavity of pillar[5]arenes and the hydrophobic fragment of the guest is located in the cavity.
- 4 April 2016
- Published by Royal Society of Chemistry (RSC) in Organic & Biomolecular Chemistry
- Vol. 14 (18) , 4233-4238
Open AccessHighly selective binding of methyl orange dye by cationic water-soluble pillar[5]arenes
A new water-soluble pillar[5]arene with an amide fragment and triethylammonium groups was synthesized by our original method of aminolysis of the ester groups. Using UV-spectroscopy, it is shown that cationic pillar[5]arenes are able to selectively form 1 : 1 complexes with some hydrophobic anions: the guests with bulky uncharged or negatively charged substituents hindering entry into the macrocycle cavity. Highly selective binding of the most lipophilic guest, methyl orange dye, in the form of organic anion salts by positively charged water-soluble pillar[5]arenes was detected. In the case of the azo dye the appropriate Kass values were 10–100-fold higher than those calculated for the other sulfonic acid derivatives studied. The 2D NMR NOESY 1H–1H spectroscopy confirms the formation of the inclusion complex: negative charge sulfonate head is outside the cavity of pillar[5]arenes and the hydrophobic fragment of the guest is located in the cavity.