GATE

GATE models and annotates genomes by detecting similar asynchronous epigenomic changes across genomic segments to infer regulatory elements and epigenomic dynamics.

Key Features:

• Probabilistic modeling: Implements a probabilistic model that detects asynchronous and spatially organized changes in epigenomic marks across genomic segments.
• Clustering Based on Temporal Changes: Clusters genomic sequences by similarity in temporal trajectories of multiple epigenomic marks during differentiation.
• High-Throughput Sequencing Integration: Integrates high-throughput sequencing measurements of seven histone modifications (H3K4me1, H3K4me2, H3K4me3, H3K27ac, H3K27me3, H3K36me3, H2A.Z), two DNA modifications (5-mC, 5-hmC), and mRNAs and noncoding RNAs.
• Predictive Capability: Predicts bidirectional promoters, miRNA promoters, and piRNAs from characteristic epigenomic patterns.
• Combinatorial-rule Derivation: Derives combinatorial rules relating combinations of epigenomic changes to mRNA and ncRNA expression levels.
• Early piRNA Signature Detection: Detects epigenomic signatures on piRNA genes that manifest prior to germ cell development.

Scientific Applications:

• Regulatory Function Investigation: Examines epigenome dynamics to infer regulatory functions and element activity.
• Epigenomic Pattern Recognition: Distinguishes gene bodies, promoters, and enhancers based on spatiotemporal epigenomic information.
• Identification of Epigenetic Signals: Identifies upstream epigenetic signals such as H3K4me2 and unmethylated CpG that correlate with TET enzyme targeting and resulting 5-hmC changes.
• Identification of Regulatory Interactions: Identifies candidate regulatory interactions and feedback loops, exemplified by Sox17 and Foxa2 during mesendoderm development.

Methodology:

Applies a probabilistic model to detect asynchronous spatial and temporal epigenomic changes and clusters genomic regions by temporal similarity; integrates high-throughput sequencing measurements of specified histone marks, DNA modifications, and transcripts; analysis was applied to mouse embryonic stem cells differentiated into mesendoderm with data collected at three time points and used to derive combinatorial epigenomic-expression rules.