METANet: a supervised ensemble learning framework for reconstructing direct and functional tissue-specific transcription factor networks.

METANet predicts tissue-specific transcription factor targets with over 30% higher accuracy than existing methods across 36 human tissues.

• Why it matters: Accurately reconstructing tissue-specific TF networks is crucial for understanding gene regulation but remains difficult due to limitations of current motif- and expression-based approaches. Improving this can enhance insights into gene regulation and disease mechanisms.
• What they did: The team developed METANet, a supervised ensemble learning framework that integrates TF motifs, cis-regulatory activity, and expression features, applying it to data from 36 human tissues to predict TF binding sites.
• The result: METANet maps outperform established methods in identifying functional TF targets validated by ChIP-seq, enable tissue-specific regulation analysis, and facilitate reproducible gene-trait association discovery.