metaLINCS: an R package for meta-level analysis of LINCS L1000 drug signatures using stratified connectivity mapping

Area of Science:

• * Computational Biology
• * Bioinformatics
• * Systems Biology

Background:

• * Perturbed gene expression profiles, like LINCS L1000, are typically analyzed individually.
• * Current methods involve summarizing data by counting individual hits per perturbagen, which can be complex.
• * A need exists for streamlined approaches to interpret large-scale gene expression datasets.

Purpose of the Study:

• * To introduce the metaLINCS R package for an alternative analysis of perturbed gene expression profiles.
• * To develop a method combining rank correlation and gene set enrichment analysis for meta-level enrichment.
• * To simplify the interpretation of large datasets and identify overarching themes.

Main Methods:

• * Utilized rank correlation and gene set enrichment analysis within the metaLINCS R package.
• * Applied the package to identify meta-level enrichment at the perturbagen and mechanism of action levels.
• * Compared the performance of metaLINCS against three existing analytical approaches.

Main Results:

• * metaLINCS effectively identifies meta-level enrichment, simplifying data interpretation.
• * The package highlights overarching themes within large gene expression datasets.
• * Demonstrated the functionality and comparative performance of the metaLINCS package.

Conclusions:

• * metaLINCS provides a powerful and simplified approach for analyzing perturbed gene expression data.
• * The package facilitates the identification of key perturbagens and their mechanisms of action.
• * metaLINCS represents a significant advancement in the analysis of large-scale biological connectivity maps.