Identification of Pou5f1, Sox2, and Nanog downstream target genes with statistical confidence by applying a novel algorithm to time course microarray and genome-wide chromatin immunoprecipitation data

Abstract
Background: Target genes of a transcription factor (TF)Pou5f1(Oct3/4orOct4), which is essential for pluripotency maintenance and self-renewal of embryonic stem (ES) cells, have previously been identified based on their response toPou5f1manipulation and occurrence of Chromatin-immunoprecipitation (ChIP)-binding sites in promoters. However, many responding genes with binding sites may not be direct targets because response may be mediated by other genes and ChIP-binding site may not be functional in terms of transcription regulation.Results: To reduce the number of false positives, we propose to separate responding genes into groups according to direction, magnitude, and time of response, and to apply the false discovery rate (FDR) criterion to each group individually. Using this novel algorithm with stringent statistical criteria (FDR < 0.2) to a compendium of published and new microarray data (3, 6, 12, and 24 hr afterPou5f1suppression) and published ChIP data, we identified 420 tentative target genes (TTGs) forPou5f1. The majority of TTGs (372) were down-regulated afterPou5f1suppression, indicating that thePou5f1functions as an activator of gene expression when it binds to promoters. Interestingly, many activated genes are potent suppressors of transcription, which include polycomb genes, zinc finger TFs, chromatin remodeling factors, and suppressors of signaling. Similar analysis showed thatSox2andNanogalso function mostly as transcription activators in cooperation withPou5f1.Conclusion: We have identified the most reliable sets of direct target genes for key pluripotency genes –Pou5f1,Sox2, andNanog, and found that they predominantly function as activators of downstream gene expression. Thus, most genes related to cell differentiation are suppressed indirectly.