piscesCSM

piscesCSM leverages graph-based representations to model small molecule chemical structures and predict drug combinations with favorable anticancer synergistic effects against one or multiple cancer cell lines.

Key Features:

• Graph-based molecular representation: Models small molecule chemical structures using graph-based representations for prediction.
• Multi-cell line prediction: Guides prediction of anticancer synergistic drug combinations in over 30 cancer cell lines.
• General supervised machine learning model: Achieves an area under the receiver operating characteristic curve (AUROC) of up to 0.89 on independent non-redundant blind tests.
• Interpretability analysis: Identifies that simple physicochemical properties and graph-based signatures are predictive of chemotherapy synergism.

Scientific Applications:

• Synergistic drug combination prediction: Prioritizes drug combinations with favorable anticancer effects for further experimental evaluation.
• Combinatorial screening library optimization: Augments screening libraries to identify effective and safe synergistic anticancer drug combinations.
• Large-scale oncology screening: Outperforms state-of-the-art approaches on large-scale oncology screening data and independent test sets.

Methodology:

piscesCSM uses graph-based representations of small molecule chemical structures combined with physicochemical properties in a general supervised machine learning model.