Skip to main content
Figure 10 | Genome Biology

Figure 10

From: A simple metric of promoter architecture robustly predicts expression breadth of human genes suggesting that most transcription factors are positive regulators

Figure 10

The clustering of the breadth of expression with the number of transcription factor binding sites. The breadth of expression clustered with RNA polymerase II and the transcription initiation factor TFIID (TAF1). More interestingly, in human tissues, the breadth of expression also clustered tightly with Mxi1, YY1, NFKB, HEY1, Sin3A, and c − Myc, suggesting these transcription factors were key in determining the breadth of expression (this cluster was marked as A). Other transcription factors formed two clusters with low and high distance to the breadth of expression (these clusters were marked as B and C, respectively). Many clusters of co-localizing transcription factors could be observed and were annotated (a)—(y). To test the robustness of this analysis, the number of transcription factors was measured in several different ways, which reassuringly clustered together and proved indistinguishable (part of the A cluster). The different measures were: sum of all sites (marked as Tfbs_x_length), sum of unique sites (Tfbs_x_length_unique), sum of all sites without RNA polymerase II (Tfbs_x_length_noPol2), and finally the sum of unique sites without the polymerase (Tfbs_x_length_unique_noPol2). The breadth of expression was also transformed in several ways which proved equivalent by forming a tight cluster (part of the A cluster). Namely, the breadth of expression was encoded as either a continuous variable (marked as breadth_continuous), discretized into three bins (breadth_discrete_3), discretized into ten bins (breadth_discrete_10), or transformed and expressed as a square root (breadth_sqrt_1).

Back to article page