A data integration framework for prediction of transcription factor targets

Area of Science:

• Computational biology
• Genomics
• Bioinformatics

Background:

• Transcription factors (TFs) are crucial proteins that regulate gene expression.
• Identifying TF targets is essential for understanding cellular processes and disease mechanisms.
• Current methods for TF target prediction have limitations in accuracy and scope.

Purpose of the Study:

• To develop and validate a computational framework for predicting transcription factor (TF) targets.
• To integrate diverse data sources, including DNA sequence and gene expression, for improved prediction accuracy.
• To demonstrate the framework's efficacy using BCL6 (B-cell lymphoma 6) as a case study.

Main Methods:

• A weighted sum approach was used to integrate multiple evidence sources.
• Evidence sources were prioritized using a training set, and their contributions were optimized.
• The framework incorporates DNA-sequence analysis and gene-expression data.
• Biological prior information was utilized to enhance prediction reliability.

Main Results:

• The computational framework successfully predicted BCL6 targets with high reliability.
• Performance significantly improved compared to methods not incorporating sequence information.
• Effective utilization of biological prior information, especially sequence analysis, was key to reliable predictions.
• The framework demonstrated a considerable gain in predictive performance.

Conclusions:

• The developed computational framework provides a reliable method for predicting transcription factor targets.
• Integrating diverse data, particularly DNA sequence information, enhances prediction accuracy.
• Careful assessment of data quality and biological relevance is crucial for successful computational predictions.