MANTRA

MANTRA analyzes drug-induced gene expression data using network theory and non-parametric statistics to map similarities between compounds and infer modes of action and potential repositioning opportunities.

Key Features:

• Network-Based Analysis: Organizes drugs as nodes in a network with edges representing similarity of drug-induced transcriptional profiles to identify communities that share modes of action.
• Non-Parametric Statistics: Applies non-parametric statistical methods to gene expression data to quantify similarity between drug-induced transcriptional profiles.
• User-Generated Data Integration: Processes uploaded pre- and post-treatment gene expression profiles across one or multiple cell types through an automated pipeline to create a node representing the specific compound.
• Neighborhood and Community Analysis for MoA Inference: Enables analysis of neighboring drugs and drug communities within the network to generate hypotheses about molecular targets, off-targets, and repositioning.

Scientific Applications:

• Identification of Molecular Targets and Off-Targets: Uses comparative analysis of drug-induced transcriptional profiles to suggest molecular targets and potential off-target effects.
• Drug Repositioning: Identifies existing drugs with shared modes of action in the network to suggest new therapeutic uses for known compounds.

Methodology:

Constructs a drug similarity network where nodes represent drugs and edges represent similarity of drug-induced transcriptional profiles computed using non-parametric statistics, with user-uploaded pre/post-treatment gene expression data processed through an automated pipeline to generate network nodes.