Past projects

Deep learning

• A simple deep learning model about adverse drug reactions - this is like an introductory piece to the world of deep learning and is also an early prototype of a QSAR/QSPR workflow as well

Machine learning

• Tree series - Decision tree 1 - data collection and preprocessing, 2 - data preprocessing and transformation, 3 - model building and estimating experimental errors, Random forest - model building, imbalanced dataset, feature importances & hyperparameter tuning, Random forest classifier - more on imbalanced dataset, Boosted trees - AdaBoost, XGBoost and Scikit-mol

• Logistic regression - 1 - Parquet file in Polars dataframe library, 2 - Preprocessing data in Polars dataframe library, 3 - Building logistic regression model using scikit-learn, 4 - Evaluating logistic regression model in scikit-learn, older long version

Data explorations

• Cytochrome P450 and small drug molecules with a focus on CYP3A4 and CYP2D6 inhibitors

• Working with scaffolds in small molecules - Manipulating SMILES strings

• Molecular similarities in selected COVID-19 antivirals - Using RDKit’s similarity map and fingerprint generator

Web applications

• Molecular visualisation web application - Interactive data table, Using Shiny for Python web application framework

• Shinylive app in Python - Data preparation, Embedding app in Quarto document & using pyodide.http