MoAn

MoAn identifies discriminative regulatory motifs in promoter sequences to map transcription factor binding sites and support gene regulatory network reconstruction.

Key Features:

• Discriminative Approach: Uses a large negative set comprising a significant sample of promoters from the relevant genome to contrast positive patterns against background and distinguish true regulatory motifs from noise.
• Probabilistic Modeling: Represents sequences with a probabilistic model and maximizes the probability of motif sets given the motif model and prior probabilities of motif occurrences in positive and negative datasets.
• Enhanced Suffix Array: Employs an enhanced suffix array to manage large negative promoter sets efficiently and improve speed and performance for large genomic datasets, including higher metazoans.
• Robustness and Accuracy: Demonstrates higher accuracy than contemporary methods, robustness to extended input sequence length, and an objective function that strongly correlates with correct motif solutions.
• Reduced Need for Repeat Masking: Reduces the necessity for repeat masking by using a large background set of real promoters, maintaining discriminatory power without separate repeat-masking preprocessing.

Scientific Applications:

• Gene Regulatory Network Mapping: Identifies regulatory elements in promoter regions to facilitate comprehensive mapping of gene regulatory networks across organisms.
• Transcription Factor Binding Site Identification: Pinpoints TFBS with high precision to aid investigation of transcriptional regulation mechanisms.

Methodology:

Leverages a large negative promoter set and probabilistic modeling to address the "pattern drowning" problem and uses an enhanced suffix array to process large datasets efficiently.