ScerTF

ScerTF catalogs over 1,200 position weight matrices (PWMs) corresponding to 196 Saccharomyces cerevisiae transcription factors to support analysis of yeast transcription factor specificities and regulatory interactions.

Key Features:

• Comprehensive PWM Collection: Aggregates motifs from 11 diverse sources into a single database covering over 1,200 PWMs for 196 yeast transcription factors.
• Benchmarking and Optimization: Rigorously benchmarks matrices against chromatin immunoprecipitation (ChIP) data and transcription factor deletion experiments to identify the most predictive matrix for each TF.
• Threshold Optimization: Optimizes score thresholds for identifying regulatory sites using in vivo data to improve prediction accuracy of TF binding.
• Matrix Combination Strategy: Combines matrices derived by different methods to mitigate methodological biases and often improve performance relative to individual PWMs.
• Prediction of Regulatory Interactions: Predicts co-occurring regulatory elements across the genome and correlates predicted interactions with gene expression data to identify TF combinations.
• Matrix Matching for TF Identification: Matches input matrices or regulatory site models to candidate TFs to aid identification of likely binding factors.

Scientific Applications:

• Linking Regulatory Sites to Transcription Factors: Assigns likely TFs to specific regulatory sequences using curated and benchmarked PWMs.
• Transcription Factor Identification: Identifies candidate TFs corresponding to a user-provided binding matrix or motif.
• Gene Regulation Analysis: Infers genes regulated by particular TFs and elucidates combinatorial regulatory interactions by integrating PWM predictions with gene expression correlations.

Methodology:

Aggregates motifs from 11 sources; benchmarks PWMs against ChIP data and transcription factor deletion experiments to select the most predictive matrix per TF; optimizes score thresholds using in vivo data; combines matrices to reduce derivation biases; predicts co-occurring regulatory elements and correlates them with gene expression data.