Evaluating the Effectiveness of Machine Learning Models for Classifying Chemical Inhibitors: A Case Study of Aromatase Inhibitors and Pubchem's Molecular Fingerprint Descriptors
Inproceedings
عنوان الكتاب :
2023 International Conference on Networking and Advanced Systems (ICNAS)
المكان :
Skikda, Algeria
الناشر :
معلومات
الفترة :October2023
الصفحات :1-7
التفاصيل
Evaluating the Effectiveness of Machine Learning Models for Classifying Chemical Inhibitors: A Case Study of Aromatase Inhibitors and Pubchem's Molecular Fingerprint Descriptors
In place of traditional experimentation, modern drug development increasingly makes use of machine learning approaches, particularly for making predictions based on the chemical structure of components. As a case study, Aramotase is a catalyst that has been linked to hormonal abnormalities, which in turn can have negative effects on sexual and skeletal maturation. Whereas, it is a major obstacle to create molecules with the aromatase enzyme. Using the ChEMBL database, which at the outset had 2818 compounds of aromatase inhibitors, we offer here a large-scale categorization structure-activity connection analysis. For further investigation, the Ic50 have been chosen as the bioactivity unit. The data curation process resulted in a final dataset consisting of 2071 compounds. We used a methodology to test the performance of different classifiers built using Pubchem's molecular fingerprint descriptors to solve the problem at hand, namely, Naive Bayes, Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors, Multilayer Perceptron, and Ad-aBoost. Considering the model's quality metrics, the Multilayer perceptron was considered the most effective model with 80 % for both accuracy and sensitivity for 10-fold cross-validation and 75% and 81 %, respectively, on the test set, and also an Fl-measure of 75%.
الكلمات المفتاحية :
Logistic regression
Sensitivity
Inhibitors
Biological system modeling
Buildings
Fingerprint recognition
Multilayer perceptrons
Machine Learning
Aromatase
Breast Cancer
QSAR
Pubchem Fingerprints
A. Abdelkrim, A. Bouramoul, I. Zenbout and S. Brahimi. 2023. Evaluating the Effectiveness of Machine Learning Models for Classifying Chemical Inhibitors: A Case Study of Aromatase Inhibitors and Pubchem's Molecular Fingerprint Descriptors. In Proceedings of the 2023 International Conference on Networking and Advanced Systems (ICNAS), Skikda, Algeria (October 2023), IEEE, 1-7. DOI: https://doi.org/10.1109/ICNAS59892.2023.10330448.
APA :
Abdelkrim, A., Bouramoul, A., Zenbout, I. & Brahimi, S. (2023, October). Evaluating the Effectiveness of Machine Learning Models for Classifying Chemical Inhibitors: A Case Study of Aromatase Inhibitors and Pubchem's Molecular Fingerprint Descriptors. In Proceedings of the 2023 International Conference on Networking and Advanced Systems (ICNAS), Skikda, Algeria, IEEE, 1-7. DOI: https://doi.org/10.1109/ICNAS59892.2023.10330448
IEEE :
A. Abdelkrim, A. Bouramoul, I. Zenbout and S. Brahimi, "Evaluating the Effectiveness of Machine Learning Models for Classifying Chemical Inhibitors: A Case Study of Aromatase Inhibitors and Pubchem's Molecular Fingerprint Descriptors". In Proceedings of the 2023 International Conference on Networking and Advanced Systems (ICNAS), Skikda, Algeria, IEEE, pp. 1-7, October, 2023. DOI: https://doi.org/10.1109/ICNAS59892.2023.10330448.
BibTeX :
@inproceedings{misc-lab-429, author = {Abdelkrim, Ali and Bouramoul, Abdelkrim and Zenbout, Imene and Brahimi, Said}, title = {Evaluating the Effectiveness of Machine Learning Models for Classifying Chemical Inhibitors: A Case Study of Aromatase Inhibitors and Pubchem's Molecular Fingerprint Descriptors}, booktitle = {2023 International Conference on Networking and Advanced Systems (ICNAS)}, location = {Skikda, Algeria}, pages = {1--7}, publisher = {IEEE}, year = {2023}, month = {October}, doi = {10.1109/ICNAS59892.2023.10330448}, url = {https://ieeexplore.ieee.org/document/10330448}, keywords = {Logistic regression, Sensitivity, Inhibitors, Biological system modeling, Buildings, Fingerprint recognition, Multilayer perceptrons, Machine Learning, Aromatase, Breast Cancer, QSAR, Pubchem Fingerprints} }
RIS :
TY - CONF TI - Evaluating the Effectiveness of Machine Learning Models for Classifying Chemical Inhibitors: A Case Study of Aromatase Inhibitors and Pubchem's Molecular Fingerprint Descriptors AU - A. Abdelkrim AU - A. Bouramoul AU - I. Zenbout AU - S. Brahimi PY - 2023 BT - 2023 International Conference on Networking and Advanced Systems (ICNAS), Skikda, Algeria SP - 1 EP - 7 PB - IEEE AB - In place of traditional experimentation, modern drug development increasingly makes use of machine learning approaches, particularly for making predictions based on the chemical structure of components. As a case study, Aramotase is a catalyst that has been linked to hormonal abnormalities, which in turn can have negative effects on sexual and skeletal maturation. Whereas, it is a major obstacle to create molecules with the aromatase enzyme. Using the ChEMBL database, which at the outset had 2818 compounds of aromatase inhibitors, we offer here a large-scale categorization structure-activity connection analysis. For further investigation, the Ic50 have been chosen as the bioactivity unit. The data curation process resulted in a final dataset consisting of 2071 compounds. We used a methodology to test the performance of different classifiers built using Pubchem's molecular fingerprint descriptors to solve the problem at hand, namely, Naive Bayes, Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors, Multilayer Perceptron, and Ad-aBoost. Considering the model's quality metrics, the Multilayer perceptron was considered the most effective model with 80 % for both accuracy and sensitivity for 10-fold cross-validation and 75% and 81 %, respectively, on the test set, and also an Fl-measure of 75%. KW - Logistic regression KW - Sensitivity KW - Inhibitors KW - Biological system modeling KW - Buildings KW - Fingerprint recognition KW - Multilayer perceptrons KW - Machine Learning KW - Aromatase KW - Breast Cancer KW - QSAR KW - Pubchem Fingerprints DO - 10.1109/ICNAS59892.2023.10330448 UR - https://ieeexplore.ieee.org/document/10330448 ID - misc-lab-429 ER -