| Conclusion | This study, and other large-scale gene expression or proteomic studies that include network and pathway analysis [82,83,114] and gene ontology [115], will help determine likely control points for cell growth, migration, and differentiation in individual tumors. |
| Data Analysis and Clustering | Gene Ontology (G0) was determined as described [34]. |
| Embedding and Cluster Analysis | Clusters were evaluated internally, based on the primary data, and externally, using PPI and gene ontology (GO) databases (SS Fig). |
| Introduction | By combining pattern recognition techniques with gene ontology (GO) and protein-protein interaction (PPI) data, we learned that clusters that contain interacting proteins are likely to indicate functional signaling pathways [34—40]. |
| Supporting Information | Gene ontology summaries of clusters. |
| Kinetics and chromatin features underlying IEG induction | (2007) [10], we constructed a set of 444 genes assigned with the Gene Ontology annotation for nucleotide binding GO:0000166 (IEA assignments were excluded). |
| The early peak signature is enriched for lEGs and signalling pathways | Gene ontology terms associated with early peak clusters that were significantly over-repre-sented in HRG treated MCF7 cells are listed in 81 Table, see also S6 Fig (analysis performed with GOrilla [22] and REVIGO [23]). |
| peak category. | Gene Ontology process term enrichment in MCF7-HRG cells. |