Multi-omics assessment of dilated cardiomyopathy using non-negative matrix factorization

Area of Science:

• Cardiovascular Biology
• Genomics
• Epigenetics

Background:

• Dilated cardiomyopathy (DCM) is a heterogeneous myocardial disease leading to heart failure and sudden cardiac death.
• Limited access to cardiac tissue has restricted the study of gene regulatory networks in DCM.
• Understanding DCM pathogenesis requires integrated multi-omics approaches.

Purpose of the Study:

• To perform an integrated analysis of transcriptome and methylome data in DCM patients.
• To uncover latent factors and covarying features between gene expression and DNA methylation.
• To identify novel interactions and biological processes involved in DCM pathogenesis.

Main Methods:

• Integrated analysis of DNA methylation (Infinium HM450) and mRNA sequencing data from DCM and control probands.
• Non-negative matrix factorization (NMF) using the R NMF package.
• Mann-Whitney U test for statistical significance and correlation network analysis.

Main Results:

• Four out of five latent factors derived from NMF were significantly different between DCM and control groups (P<0.05).
• Enrichment analysis revealed significant involvement of immune response, nucleic acid binding, extracellular matrix, and myofibrillar structure pathways in DCM.
• Identified novel correlations between sarcomeric genes and CpG methylation sites, including ATPase Phospholipid Transporting 11A0 and Solute Carrier Family 12 Member 7.

Conclusions:

• Matrix factorization enables integration of multi-omics data from human tissues to identify novel interactions in complex diseases like DCM.
• This hypothesis-generating approach can enhance understanding of DCM pathophysiology.
• The findings highlight potential gene-environment interactions in DCM development.