This repository provides essential tools and example workflows to support participants in the Target2035 challenge. It includes utilities for molecular fingerprint extraction, data loading, model training, and evaluation.
-
Fingerprint Extraction
Scripts to compute molecular fingerprints (ECFP4, ECFP6, MACCS, etc.) from SMILES and save them into a Parquet file.
📁 folder name: FingerprintExtraction -
Parquet Reader Utilities
Simple tools to load Parquet files and process fingerprint columns into NumPy arrays.
📁 folder name: ReadingParquetFiles -
Training Notebook
A notebook covering training set loading, cross-validation, model training, and optional clustering of results.
📁 folder name: Notebooks -
Evaluation Code
Scripts that accept the gold label files and participants' predicted outputs, and compute evaluation metrics.
📁 folder name: EvaluationCode