CombSAFE

CombSAFE identifies and analyzes combinations of static and dynamic functional elements across the genome or within specific genomic regions under semantically annotated biological conditions.

Key Features:

• Integration of heterogeneous ChIP-seq data: Integrates heterogeneous ChIP-seq datasets from public repositories and associates samples with semantic annotations using natural language processing and selected biomedical ontologies, enabling cross-condition analysis of DNA-associated protein binding sites including transcription factors and histone marks.
• Identification of functional element combinations: Uses hidden Markov models to detect combinations of static and dynamic functional genomic elements across the genome.
• Genome-wide analysis and clustering: Performs genome-wide clustering of regions with similar patterns of functional elements and conducts enrichment analyses to identify ontological terms significantly associated with those clusters.
• Comparative functional state analysis: Compares functional states of specific genomic regions across different semantic annotations to reveal condition-specific regulatory patterns and unexpected combinations of functional elements.

Scientific Applications:

• Gene regulation analysis: Dissect combinatorial arrangements of transcription factor binding and histone modification patterns to investigate gene regulatory mechanisms across conditions.
• Epigenetic state characterization: Characterize epigenetic modifications and their dynamics by integrating histone mark ChIP-seq data with semantic context.
• Comparative condition and disease studies: Compare functional genomic states across biological conditions to investigate implications in health and disease.
• Discovery of regulatory networks: Reveal novel regulatory networks by identifying unexpected combinations of functional elements across ontologically annotated datasets.

Methodology:

Integration of ChIP-seq data with semantic annotations using natural language processing and biomedical ontologies, application of hidden Markov models to detect combinations of functional elements, clustering of genomic regions by element patterns, enrichment analysis for ontological term association, and comparative analysis of functional states across semantic annotations.