Multiscale representation of genomic signals

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Genomic information exists across diverse distance scales, from base pairs to megabases.
• Understanding signal patterns at multiple scales is crucial for biological interpretation.
• Existing methods often focus on single scales, potentially missing broader genomic context.

Purpose of the Study:

• To develop and present a multiscale framework for analyzing genomic signal information content.
• To associate multiscale signal patterns with genomic annotations and biological processes.
• To enhance the prediction of gene expression and investigate epigenetic modifications in cancer.

Main Methods:

• Development of a multiscale analytical framework for genomic signals.
• Characterization of signal enrichment/depletion patterns across various scales (e.g., G+C content, DNA methylation).
• Integration of multiscale information for predictive modeling and association studies.

Main Results:

• Distinct multiscale patterns were identified for different genomic signals like G+C content and DNA methylation.
• These patterns were successfully associated with various genomic annotations.
• Improved prediction of gene expression was achieved by integrating information across all scales.
• Methylation patterns extending beyond single-gene scales were linked to gene expression differences in colorectal cancer.

Conclusions:

• A multiscale approach provides a more comprehensive understanding of genomic information.
• The framework enhances predictive capabilities for gene expression and reveals novel insights into epigenetic regulation.
• The developed software facilitates the analysis of genomic data across multiple scales.