Abasy Atlas

Abasy Atlas provides meta-curated reconstructed bacterial gene regulatory networks (GRNs) and analytical data to support system-level, cross-species, and benchmarking studies of bacterial gene regulation.

Key Features:

• Meta-curated reconstructed GRNs: Consolidated and curated GRNs compiled from multiple sources for high-quality cross-organism analyses.
• Harmonization of datasets and naming conventions: Reconciles inconsistent gene names and network representations to reduce duplicated interactions and biased analyses.
• Historical snapshots of regulatory networks: Retains temporal versions of GRNs to perform analyses at different levels of completeness and identify biases over time.
• Interaction-count model (genome-size dependent): Quantifies the total number of regulatory interactions as a function of genome size to inform completeness estimation.
• Dataset scope and composition: Contains 76 networks with 204,282 regulatory interactions across 42 bacterial species, with 64% Gram-positive and 36% Gram-negative representatives including Mycobacterium tuberculosis, Bacillus subtilis, Escherichia coli, Corynebacterium glutamicum, Staphylococcus aureus, Pseudomonas aeruginosa, Streptococcus pyogenes, Streptococcus pneumoniae, and Streptomyces coelicolor.
• Regulons and modules: Includes 8,459 regulons and 4,335 modules.

Scientific Applications:

• System-level GRN analysis: Supports analyses of GRN architecture, organization, and system-level properties in bacteria.
• Cross-species comparative studies: Enables comparative analyses of regulatory networks across multiple bacterial species.
• Benchmarking and gold standards: Provides meta-curated networks suitable for benchmarking inference methods and establishing gold-standard datasets.
• Network completeness estimation and curation: Uses the interaction-count model and historical snapshots to guide curation, prediction, and validation of GRNs.
• Temporal bias identification and prediction: Employs historical snapshots to identify temporal biases and predict future network properties.

Methodology:

Reconstruction and meta-curation of GRNs, harmonization of gene names and network representations to remove duplicated interactions, provision of historical snapshots of networks, and a quantitative model estimating total regulatory interactions as a function of genome size.