Transcription factor co-localization patterns affect human cell type-specific gene expression.

Area of Science:

• Molecular Biology
• Genomics
• Systems Biology

Background:

• Precise control of gene expression states is fundamental for cellular development.
• Transcription factors regulate gene expression through combinatorial DNA binding.
• The transferability of transcription factor binding data across different cell types remains largely unknown.

Purpose of the Study:

• To investigate the extent to which transcription factor binding data from one cell type can predict gene regulation in other cell types.
• To develop regression models linking co-localized transcription factor binding sites to target gene expression.
• To assess the importance of cell-type specific binding patterns in understanding combinatorial gene regulation.

Main Methods:

• Integrated transcription factor co-localization data from K562 cells with gene expression data from 38 hematopoietic cell types.
• Developed regression models to correlate transcription factor binding with target gene expression.
• Compared prediction accuracy using binding data from K562 and GM12878 cell lines.

Main Results:

• Regression models significantly predicted gene expression only in monocytic cells, closely related to K562 (p<0.001).
• Cell-type specific binding patterns were critical for accurate prediction.
• Differences in binding patterns between K562 and GM12878 correlated with prediction quality.
• Successfully linked transcription factors FOS, TAF1, and YY1 to the gene LRIG2.
• Identified cell-type and co-factor dependent variations in transcription factor activity.

Conclusions:

• Gene expression can be explained by a limited set of co-localized transcription factors.
• Cell-type specific binding information is essential for accurate prediction of combinatorial gene regulation.
• The study highlights the critical role of context-dependent transcription factor interactions.