Promzea

Promzea predicts and prioritizes cis-regulatory motifs in plant upstream sequences by integrating outputs from BioProspector, Weeder, and MEME to support discovery of regulatory networks in maize, rice, and Arabidopsis.

Key Features:

• Integration of Multiple Algorithms: Promzea leverages BioProspector, Weeder, and MEME and combines their distinct motif outputs to enhance motif prediction accuracy.
• Optimized Statistical Filters: Statistical filters were optimized on a benchmark dataset to reduce the false discovery ratio and improve prediction reliability.
• Upstream Sequence Retrieval: The system accepts cDNA sequences or gene IDs and retrieves corresponding upstream sequences from the maize, rice, or Arabidopsis genomes.
• Motif Filtering, Combination, and Ranking: Predicted motifs are filtered, combined, and ranked according to relevance and likelihood of biological significance.
• Genome-wide Motif Searches and Annotation: Promzea conducts genome-wide searches for genes containing each predicted motif and returns lists of genes with their annotations.

Scientific Applications:

• Improved Sensitivity: Promzea achieved a 22% increase in nucleotide sensitivity compared to Weeder while maintaining equivalent specificity.
• In Silico Benchmarking: The approach retrieved benchmark motifs and experimentally defined binding sites for transcription factors in the maize anthocyanin and phlobaphene biosynthetic pathways.
• Discovery of Broader Regulatory Networks: Promzea identified 127 non-anthocyanin/phlobaphene genes containing all five predicted promoter motifs, indicating broader co-regulation.
• Tissue-Specific Analysis: The method was tested against tissue-specific maize microarray data to validate its utility for uncovering complex, tissue-associated regulatory networks.

Methodology:

Integration of outputs from BioProspector, Weeder, and MEME; optimization of statistical filters using a benchmark dataset; retrieval of upstream sequences from maize, rice, and Arabidopsis given cDNA sequences or gene IDs; filtering, combining, and ranking of predicted motifs; genome-wide searches for motif occurrences with retrieval of gene annotations.