MAPPER Database

MAPPER Database catalogs predicted transcription factor binding sites across human, mouse, Drosophila melanogaster, Caenorhabditis elegans, and Saccharomyces cerevisiae genomes to support comparative analysis of cis-regulatory elements.

Key Features:

• Extensive Database: MAPPER Database contains predicted binding sites in promoter regions across multiple model organisms represented by over 600 transcription factors and 1,017 probabilistic models derived from experimentally determined TFBSs in TRANSFAC and JASPAR.
• Advanced Search Methodology: The database uses hidden Markov models constructed from alignments of known binding sites to detect putative regulatory elements with reported higher sensitivity and specificity than nucleotide weight matrix-based approaches in experimental and synthetic-data comparisons.
• Multi-Organism Analysis: It supports a multi-organism catalog of annotated TFBSs with homology information across clusters of orthologous genes for comparative studies.
• Model Building and Sequence Querying: The platform enables building models from multiple sequence alignments of binding sites and querying uploaded sequences against those models.
• Database Expansion with Experimental Data: The resource incorporates experimentally determined TFBS data to extend model coverage and analysis capabilities.

Scientific Applications:

• Cis-regulatory module analysis: MAPPER Database facilitates identification and analysis of combinations of regulatory elements that control spatial and temporal gene expression.
• Transcriptional regulation research: It supports studies of transcription factor binding patterns underlying gene regulation in development and disease.
• Comparative and evolutionary genomics: The multi-species TFBS catalog enables comparison of conserved and species-specific transcription factor binding across genomes.
• Promoter-focused regulatory mapping: It enables mapping predicted TFBSs in promoter regions to link regulatory motifs to gene targets.

Methodology:

Predictions are generated using hidden Markov models constructed from alignments of known binding sites; 1,017 probabilistic models are derived from experimentally determined TFBSs sourced from TRANSFAC and JASPAR; model building uses multiple sequence alignments of binding sites; performance has been compared to nucleotide weight matrix-based approaches using experimental validations and synthetic-data comparisons; annotated TFBSs are organized into multi-organism catalogs with homology information across clusters of orthologous genes.