I went for a deeper dive this time looking at building a decision tree... it ended up with:
Post 1 - data collection from ChEMBL database using web resource client (https://jhylin.github.io/Data_in_life_blog/posts/16_ML2-1_Decision_tree/1_data_col_prep.html)
Post 2 - more data preprocessing and transformation (https://jhylin.github.io/Data_in_life_blog/posts/16_ML2-1_Decision_tree/2_data_prep_tran.html)
Post 3 - building decision tree model using scikit-learn & estimating experimental errors (https://jhylin.github.io/Data_in_life_blog/posts/16_ML2-1_Decision_tree/3_model_build.html)
#ml #decisiontree #scikitlearn #pandas #rdkit #datamol #cheminformatics #seaborn #graphviz #dtreeviz #python
Here's another web app built aiming towards non-coding lab chemists (my two cents towards this) - https://jhylin.github.io/Data_in_life_blog/posts/15_Molviz/Molviz.html
Quick features of the app:
- viewing and highlighting substructures in 2D compound images (for deployed version)
- viewing, saving and highlighting 2D images (for localhost version)
For interactive data table part: https://jhylin.github.io/Data_in_life_blog/posts/15_Molviz/itables.html
Big thanks to #rdkit, #datamol, #pyshiny, #itables, #pandas, #polars & #python.
Latest blog post: Working with scaffolds in small molecules - manipulating SMILES strings with functions to read & query multiple SMILES. Many thanks to datamol, RDKit and several other great work with acknowledgements mentioned in post - https://jhylin.github.io/Data_in_life_blog/posts/14_Scaffolds_in_small_molecules/chembl_anti-inf_data_prep_current.html
#drugdiscovery #datamol #rdkit #compchem #cheminformatics
