RNA Solutions: Synthesizing Information to Support Transcriptomics (RNASSIST).

Area of Science:

• Genomics
• Computational Biology
• Neuroscience

Background:

• Transcriptomics is vital for disease gene discovery.
• Standard analysis of differentially expressed genes (DEGs) often misses subtle signals, especially in tissues like the brain.
• Identifying relevant genes with small effect sizes remains a challenge in transcriptomic research.

Purpose of the Study:

• To develop a novel computational approach, RNASSIST, to identify hidden signals in transcriptomic data.
• To improve the detection of disease-associated genes, particularly those with small effect sizes.
• To apply RNASSIST to RNA-sequencing data from post-mortem brains to study Alcohol Use Disorder (AUD).

Main Methods:

• Developed RNASSIST, a machine learning algorithm integrating differential expression and co-expression network analysis.
• Applied RNASSIST to RNA-sequencing data comparing AUD and control post-mortem brain samples.
• Employed multiple validation strategies to confirm the biological relevance of identified genes.

Main Results:

• RNASSIST identified candidate genes not detected by standard differential expression analysis.
• Many identified genes, though not differentially expressed, were validated as potentially significant in AUD.
• The approach successfully uncovered hidden transcriptomic signals relevant to AUD.

Conclusions:

• RNASSIST effectively identifies biologically relevant genes missed by conventional transcriptomic analyses.
• The method provides a powerful tool for dissecting complex diseases like AUD.
• The identified genes offer new avenues for understanding AUD pathogenesis.