Östrojen Reseptör Alfa (ERα) İnhibitörlerinin Aktivitelerinin İstatistiksel Öğrenme ile Tahmini

Abstract

Östrojen reseptör alfa (ERα), hücre büyümesi ve çoğalması gibi süreçlerde görev alan bir proteindir; ancak meme kanserlerinin %70'inde aşırı bulunması nedeniyle önemli bir araştırma konusudur. ERα inhibitörleri bu proteinin aktivitesini engelleyerek kanser hücrelerinin büyümesini durdurur. Geleneksel ilaç keşif yöntemleri zaman ve maliyet açısından dezavantajlıdır. ERα inhibitörlerinin keşfi için literatürde çeşitli yaklaşımlar mevcuttur. Bu nedenle, çalışmada ERα inhibitörlerinin keşfi için makine öğrenmesi tabanlı Kantitatif Yapı-Aktivite İlişkisi (Quantitative Structure-Activity Relationship - QSAR) yaklaşımı tercih edilmiştir. Bu çalışmada, geniş kimyasal veri setlerinden yapı-aktivite ilişkilerini çıkarma kapasitesi ve yüksek verimli tarama imkanı sunması nedeniyle makine öğrenmesi tabanlı QSAR yaklaşımı tercih edilmiştir. Bu yöntem, moleküllerin kimyasal yapısal özelliklerini sayısal tanımlayıcılarla ifade ederek biyolojik aktivitelerini tahmin etmeye olanak sağlar. Veri kaynağı olarak, yaygın kullanımı, yüksek veri kalitesi ve önceki çalışmalarla karşılaştırma imkanı sunması nedeniyle ChEMBL206 hedef tanımlayıcısı seçilmiştir. Elde edilen moleküller IC50 değerlerine göre sınıflandırılmış, Lipinski kurallarıyla kimyasal uzay dağılımları analiz edilmiştir. Ardından PADEL programı ile 3053 molekül için 3153 moleküler tanımlayıcı hesaplanmıştır. Özellik önemi analizinde, PubchemFP667 ve PubchemFP527 gibi parmak izlerinin ve LightGBM modelinde öne çıkan APC2D atom çifti tanımlayıcılarının ERα engellemesinde kritik rol oynadığı görülmüştür. Geliştirilen modeller %94'ün üzerinde doğruluk, %90 civarında duyarlılık, %95'in üzerinde özgüllük ve 0.97'nin üzerinde AUC değeriyle üstün performans göstermiştir. Bu çalışma, yüksek doğruluk oranıyla ERα inhibitörlerinin aktivitesinin tahmin edilebileceğini göstererek ilaç keşif sürecinin verimliliğine katkı sağlamaktadır.

Keywords

• Östrojen reseptör alfa (ERα)
• Sınıflandırma tabanlı QSAR
• Makine öğrenmesi
• Özellik önemi
• İnhibitör aktivitesi tahmini

Statistical Learning-Based Prediction of Estrogen Receptor Alpha (ERα) Inhibitor Activities

Abstract

Estrogen receptor alpha (ERα) is a protein that plays a role in processes such as cell growth and proliferation; however, it has become an important research topic due to its overexpression in 70% of breast cancers. ERα inhibitors stop the growth of cancer cells by blocking the activity of this protein. Traditional drug discovery methods are disadvantageous in terms of time and cost. Various approaches exist in the literature for the discovery of ERα inhibitors. Therefore, a machine learning-based Quantitative Structure-Activity Relationship (QSAR) approach was preferred in this study for the discovery of ERα inhibitors. In this study, a machine learning-based QSAR approach was preferred due to its capacity to extract structure-activity relationships from large chemical datasets and its ability to provide high-throughput screening opportunities. This method enables the prediction of biological activities by expressing the chemical structural properties of molecules through numerical descriptors. The ChEMBL206 target identifier was selected as the data source due to its widespread use, high data quality, and the opportunity for comparison with previous studies. The obtained molecules were classified according to their IC50 values, and their chemical space distributions were analyzed using Lipinski rules. Subsequently, 3153 molecular descriptors were calculated for 3053 molecules using the PADEL program. Feature importance analysis revealed that fingerprints such as PubchemFP667 and PubchemFP527, as well as APC2D atom pair descriptors that stood out in the LightGBM model, played critical roles in ERα inhibition. The developed models demonstrated superior performance with accuracy above 94%, sensitivity around 90%, specificity above 95%, and AUC values above 0.97. This study contributes to the efficiency of the drug discovery process by demonstrating that the activity of ERα inhibitors can be predicted with high accuracy rates.

Keywords

• Estrogen receptor alpha (ERα)
• Classification based-QSAR
• Machine learning
• Feature importance
• Inhibitor activity prediction

References

1. Al-Thanoon, N. A., Qasim, O. S., Algamal, Z. Y., 2019. A new hybrid firefly algorithm and particle swarm optimization for tuning parameter estimation in penalized support vector machine with application in chemometrics. Chemometrics and Intelligent Laboratory Systems, 184, 142–152.
2. Ali, S., Coombes, R. C., 2002. Endocrine-responsive breast cancer and strategies for combating resistance. Nature Reviews Cancer, 2(2), 101–112.
3. Anderson, E., 2002. The role of oestrogen and progesterone receptors in human mammary development and tumorigenesis. Breast Cancer Research, 4(5), 197–201.
4. Arciniegas, F., Bennett, K., Breneman, C., Embrechts, M. J., 2000. Molecular database mining using self-organizing maps for the design of novel pharmaceuticals. In Intelligent Engineering Systems through Artificial Neural Networks: Smart Engineering System Design, 10, 477–481.
5. Breiman, L., 1996. Bagging predictors. Machine Learning, 24(2), 123–140.
6. Breiman, L., 2001. Random forests. Machine Learning, 45(1), 5–32.
7. Byvatov, E., Fechner, U., Sadowski, J., Schneider, G., 2003. Comparison of support vector machine and artificial neural network systems for drug/nondrug classification. Journal of Chemical Information and Computer Sciences, 43(6), 1082–1089.
8. Carracedo-Reboredo, P., Liñares-Blanco, J., Rodríguez- Fernández, N., Cedrón, F., Novoa, F. J., Carballal, A., Maojo, V., Pazos, A., Fernandez-Lozano, C., 2021. A review on machine learning approaches and trends in drug discovery. Computational and Structural Biotechnology Journal, 19, 4538-4558.

9. Chen, T., Guestrin, C., 2016. XGBoost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785–794.
10. Deroo, B. J., Korach, K. S., 2006. Estrogen receptors and human disease. Journal of Clinical Investigation, 116(3), 561–570.
11. DiMasi, J. A., Grabowski, H. G., Hansen, R. W., 2016. Innovation in the pharmaceutical industry: New estimates of R&D costs. Journal of Health Economics, 47, 20-33.
12. Gaulton, A., Hersey, A., Nowotka, M., Bento, A. P., Chambers, J., Mendez, D., Leach, A. R., 2017. The ChEMBL database in 2017. Nucleic Acids Research, 45(D1), D945–D954.
13. Geurts, P., Ernst, D., Wehenkel, L., 2006. Extremely randomized trees. Machine Learning, 63(1), 3–42.
14. Gümüştaş, E. and Çakmak Pehlivanlı, A.Ç., 2021. In-silico mutajenisite tahmininde istatistiksel öğrenme modeli. Süleyman Demirel Üniversitesi Fen Bilimleri Enstitüsü Dergisi, 25(2), 365–370.
15. Huang, P., Chandra, V., Rastinejad, F., 2010. Structural overview of the nuclear receptor superfamily: insights into physiology and therapeutics. Annual Review of Physiology, 72, 247–272.
16. Jensen, E. V., Jordan, V. C., 2003. The estrogen receptor: a model for molecular medicine. Clinical Cancer Research, 9(6), 1980–1989.
17. Jordan, V. C., 2003. Targeting anti-hormone resistance in breast cancer: a simple solution. Annals of Oncology, 14(7), 969–970.
18. Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., Liu, T. Y., 2017. LightGBM: A highly efficient gradient boosting decision tree. Advances in Neural Information Processing Systems (NIPS 2017), 30, 3146-3154.
19. Krippendorff, B. F., Lienau, P., Reichel, A., Huisinga, W., 2007. Optimizing classification of drug-drug interaction potential for CYP450 isoenzyme inhibition assays in early drug discovery. Journal of Biomolecular Screening, 12(1), 92–99.
20. LeCun, Y., Bengio, Y., Hinton, G., 2015. Deep learning. Nature, 521, 436-444.
21. Lipinski, C. A., 2004. Lead- and drug-like compounds: the rule-of-five revolution. Drug Discovery Today: Technologies, 1(4), 337-341.
22. Lipinski, C. A., Lombardo, F., Dominy, B. W., Feeney, P. J., 2001. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Advanced Drug Delivery Reviews, 46(1-3), 3–26.
23. Musgrove, E. A., Sutherland, R. L., 2009. Biological determinants of endocrine resistance in breast cancer. Nature Reviews Cancer, 9(9), 631–643.
24. Qasim, O. S., Algamal, Z. Y., 2018. Feature selection using particle swarm optimization-based logistic regression model. Chemometrics and Intelligent Laboratory Systems, 182, 41-46.
25. Sakri, S. B., Abdul Rashid, N. B., Muhammad Zain, Z., 2018. Particle swarm optimization feature selection for breast cancer recurrence prediction. IEEE Access, 6, 29637-29647.
26. Sebaugh, J. L., 2011. Guidelines for accurate EC50/IC50 estimation. Pharmaceutical Statistics, 10(2), 128–134.
27. Shaker, B., Ahmad, S., Lee, J., Jung, C., Na, D., 2021. In silico methods and tools for drug discovery. Computational Biology in Medicine, 137, 104851.
28. Sliwoski, G., Kothiwale, S., Meiler, J., Lowe, E. W., 2013. Computational methods in drug discovery. Pharmacological Reviews, 66(1), 334–395.
29. Sokolova, M., Japkowicz, N. & Szpakowicz, S., 2006. Beyond Accuracy, F Score and ROC: A Family of Discriminant Measures for Performance Evaluation. In: A. Sattar & B. Kang, eds., AI 2006: Advances in Artificial Intelligence, Lecture Notes in Computer Science, vol. 4304, Springer, Berlin Heidelberg, 1015–1021.
30. Wagener, M., Van Geerestein, V. J., 2000. Potential drug and non-drugs: prediction and identification of important structural features. Journal of Chemical Information and Computer Sciences, 40(2), 280–292.
31. Yap, C. W., 2011. PaDEL-descriptor: An open-source software to calculate molecular descriptors and fingerprints. Journal of Computational Chemistry, 32(7), 1466–1474.
32. Zhang, J. H., Chung, T. D., Oldenburg, K. R., 2000. Confirmation of primary active substances from high throughput screening of chemical and biological populations: a statistical approach and practical considerations. Journal of Combinatorial Chemistry, 2(3), 258–265.