POCO

POCO identifies over-represented nucleotide patterns within promoter regions of co-expressed genes to detect potential transcription factor binding elements and infer shared regulatory signals.

Key Features:

• Pattern discovery: Systematically searches for nucleotide patterns within promoter sets of co-expressed genes and can analyze a single set or compare two distinct sets to find over-represented motifs.
• Comparative analysis: Evaluates promoter sets derived from microarray-defined up-regulated and down-regulated gene groups to identify patterns enriched in one set and depleted in another.
• Hypothesis testing: Tests differential pattern distribution across up-regulated, down-regulated, and randomly chosen promoter sets to implicate potential transcription factor binding elements.
• Comprehensive pattern enumeration: Performs an exhaustive search over all possible nucleotide patterns using the DNA alphabet A, C, G, T and the degenerate symbol N to represent any nucleotide.
• Statistical analysis: Uses bootstrapping to estimate mean occurrences and standard deviations of patterns, applies ANOVA F-statistics to detect differences across promoter sets, and employs Tukey's honestly significantly different (HSD) test and P-values for post-hoc significance assessment.

Scientific Applications:

• Transcription factor identification: Identifies candidate transcription factor binding elements associated with co-expressed gene sets.
• Regulatory network inference: Supports inference of regulatory interactions by linking enriched promoter motifs to co-expression patterns.
• Functional genomics: Aids interpretation of microarray-derived gene clusters to reveal promoter-level regulatory signals underlying biological responses.
• Comparative genomics: Enables comparison of motif distributions across gene sets to study conservation or divergence of regulatory elements.
• Systems biology: Provides motif-level information for modeling gene regulation within systems-level analyses.

Methodology:

Exhaustive enumeration of nucleotide patterns over the DNA alphabet (A, C, G, T, N); counting pattern occurrences in promoter sets (single or pairwise comparisons); bootstrapping to estimate mean and standard deviation of pattern counts; ANOVA F-statistics to test differences across up-regulated, down-regulated, and random promoter sets; followed by Tukey's honestly significantly different (HSD) test and P-value assessment for significance.