Statistical integration of multi-omics and drug screening data from cell lines

Area of Science:

• Computational biology
• Genomics
• Neuroscience

Background:

• Synucleinopathies, including Parkinson's disease, are complex neurological disorders.
• Current multi-omics data integration methods struggle to effectively analyze diverse datasets from cell line models.
• Identifying druggable targets for synucleinopathies remains a significant challenge.

Purpose of the Study:

• To propose a novel computational workflow for joint analysis of multi-omics data from cell lines.
• To identify potentially druggable pathways and genes involved in synucleinopathies using a new probabilistic data integration method.
• To prioritize therapeutic targets for synucleinopathies.

Main Methods:

• Development of POPLS-DA (probabilistic data integration) for multi-omics analysis.
• Integration of transcriptomics, proteomics, and drug screening data from LUHMES cell lines.
• Construction of an integrated interaction network incorporating drug screening data.

Main Results:

• POPLS-DA prioritized genes and proteins distinguishing cases and controls in synucleinopathies.
• The workflow highlighted druggable genes and pathways, identifying HSPA5 as a key target.
• Functional enrichment analysis revealed synaptic and lysosome-related gene clusters targeted by protective drugs, including AT1-blockers.

Conclusions:

• The computational workflow offers a powerful approach for analyzing multi-omics data in synucleinopathies.
• HSPA5 and AT1-blockers represent promising therapeutic avenues for synucleinopathies and Parkinson's disease.
• POPLS-DA outperforms other methods in providing a larger, interpretable gene set for target identification.