KNeMAP: a network mapping approach for knowledge-driven comparison of transcriptomic profiles.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Transcriptomic data analysis is crucial for understanding chemical compound mechanisms of action (MOA).
• Omics data complexity and noise present challenges for accurate dataset comparison.
• Existing methods comparing gene expression values or sets of differentially expressed genes can be affected by technical/biological variance and neglect gene relationships.

Purpose of the Study:

• To introduce KNeMAP (Knowledge-driven Network MApping of Profiles), a novel network mapping approach for comparing transcriptomic profiles.
• To enhance the accuracy and robustness of transcriptomic data comparison, particularly for MOA studies.
• To provide a higher-level view of transcriptomic data by grouping genes based on prior information.

Main Methods:

• KNeMAP combines genes into similarity groups using multiple levels of prior information.
• The approach integrates network mapping for knowledge-driven comparison of transcriptomic profiles.
• Performance was evaluated against fold change and gene set-based methods.

Main Results:

• KNeMAP demonstrated higher accuracy in grouping compounds compared to existing methods.
• The approach is less susceptible to noise-corrupted data.
• KNeMAP successfully identified similar molecular responses across different cell lines and datasets, including the Connectivity Map and Fortino et al. datasets.

Conclusions:

• KNeMAP offers a robust and accurate method for comparing transcriptomic profiles.
• The approach enhances the understanding of compound MOA by overcoming limitations of traditional methods.
• KNeMAP provides a valuable tool for analyzing complex omics data in various biological systems.