Materials to “OCHEM - platform for winning Challenges!” by Igor Tetko
https://opentox.net/events/opentox-summer-school-2025/Igor-Tetko
Instructions to develop models:
- Create account at OCHEM https://ochem.eu
- Upload data to https://files.ochem.eu/school/train.xls, https://files.ochem.eu/school/validation.xls and https://files.ochem.eu/school/test.xls to baskets
- Develop models using training/validation sets, and select the best one
- Predict the test set compounds
Yo can see presentation from the last year https://register.gotowebinar.com/recording/7244633078930412973 as well as short tutorials at https://ochem.eu/static/tutorials.do
During the talk the data for test set will be released and the best model prepared by each participant will be determined. You can learn about the best strategies to develop models by reading these articles:
(1) Mousa, N.; Varbanov, H. P.; Kaipanchery, V.; Gabano, E.; Ravera, M.; Toropov, A. A.; Charochkina, L.; Menezes, F.; Godin, G.; Tetko, I. V. Online OCHEM Multi-Task Model for Solubility and Lipophilicity Prediction of Platinum Complexes. J. Inorg. Biochem. 2025, 269, 112890. https://doi.org/10.1016/j.jinorgbio.2025.112890.
(2) Hunklinger, A.; Hartog, P.; Šícho, M.; Godin, G.; Tetko, I. V. The openOCHEM Consensus Model Is the Best-Performing Open-Source Predictive Model in the First EUOS/SLAS Joint Compound Solubility Challenge. SLAS Discov. 2024, 29 (2), 100144. https://doi.org/10.1016/j.slasd.2024.01.005.
(3) Eytcheson, S. A.; Tetko, I. V. Which Modern AI Methods Provide Accurate Predictions of Toxicological Endpoints? Analysis of Tox24 Challenge Results. ChemRxiv January 10, 2025. https://doi.org/10.26434/chemrxiv-2025-7k7x3.
(4) Cirino, T.; Pinto, L.; Iwan, M.; Dougha, A.; Lučić, B.; Kraljević, A.; Navoyan, Z.; Tevosyan, A.; Yeghiazaryan, H.; Khondkaryan, L.; Abelyan, N.; Atoyan, V.; Babayan, N.; Iwashita, Y.; Kimura, K.; Komasaka, T.; Shishido, K.; Nakamura, T.; Asada, M.; Jain, S.; Zakharov, A. V.; Wang, H.; Liu, W.; Chupakhin, V.; Uesawa, Y. Consensus Modeling Strategies for Predicting Transthyretin Binding Affinity from Tox24 Challenge Data. Chem. Res. Toxicol. 2025, 38 (6), 1061–1071. https://doi.org/10.1021/acs.chemrestox.5c00018.
(5) Makarov, D. M.; Ksenofontov, A. A.; Budkov, Y. A. Consensus Modeling for Predicting Chemical Binding to Transthyretin as the Winning Solution of the Tox24 Challenge. Chem. Res. Toxicol. 2025, 38 (3), 392–399. https://doi.org/10.1021/acs.chemrestox.4c00421.
(6) Han, Z.; Xia, Z.; Xia, J.; Tetko, I. V.; Wu, S. The State-of-the-Art Machine Learning Model for Plasma Protein Binding Prediction: Computational Modeling with OCHEM and Experimental Validation. Eur. J. Pharm. Sci. 2025, 204, 106946. https://doi.org/10.1016/j.ejps.2024.106946.