GMD

GMD measures non-parametric dissimilarities between discrete transcription start site (TSS) frequency distributions to classify promoter TSS distribution types and analyze their biological implications.

Key Features:

• Non-Parametric Dissimilarity Measure: Implements a non-parametric approach to quantify dissimilarities between TSS distributions (TSSDs) without assuming underlying distributional forms.
• Clustering Approach: Applies an advanced clustering methodology to categorize and analyze clusters of TSSDs by similarity and stability, extending beyond binary broad versus sharp promoter classifications.
• Identification of Ultra-Sharp TSSD Promoters: Detects ultra-sharp TSSD promoters characterized by transcription initiation at nearly identical genomic positions.
• Biological Implications: Differentiates TSSD types to inform regulatory and functional roles of core promoters and their impact on gene expression patterns.
• Error Mitigation: Identifies mapping errors associated with ribosomal protein pseudogenes and filters potential artifact ultra-sharp TSS distributions to reduce confounding in downstream analyses.

Scientific Applications:

• Gene Regulation Studies: Enables investigation of how distinct TSSD patterns affect promoter function and gene regulatory mechanisms.
• Functional Genomics: Supports characterization of core promoter diversity and stability across genomes using genome-wide TSS data.
• Epigenetic Analysis: Facilitates identification of promoters lacking typical epigenetic signatures to explore non-canonical regulatory elements.

Methodology:

Uses high-throughput genome-wide TSS detection data, a non-parametric dissimilarity measure for TSSDs, and clustering to explore and classify TSS distribution types.