Collaboration networks have proven informative when used to describe various kinds of human relationships. Similar strategy could be used in the transcriptional regulatory network. A collaboration network of target genes (TGs) was constructed based on common transcription factors (TFs), and similarly, a smaller network of transcription factors was constructed based on their common target genes. After clustering the target gene collaboration networks, genes in the same cluster were often enriched for one or more GO terms. The results also showed that genes with specific GO terms tend to share similar regulatory mechanisms. It indicates that in a collaboration network approach the relatively simple "regulatory mechanism" measure used here was able to extract considerable biologically relevant information. Moreover, a definition of anomaly used before in a bipartite graph analysis method was applied into the collaboration networks analysis. And the correlation between the anomalies and the essential genes was discovered. In a conclusion, a collaboration network approach may be a valuable supplement to other analyses of transcriptional networks.