Gene Ontology analysis was conducted for each pattern based on the EBI database. As shown in Table 1, the highenrichment GO terms included: metabolic process, WZ8040 EGFR/HER2 inhibitor response to stimuli, immune response and cellular component organization. Detailed functional category information for each temporal expression cluster is shown in Table S3. Further comprehensive GO analysis was applied to differentially expressed genes to gain deeper insight into the main cellular functions that are altered during IR recovery. This was performed according to Gene Ontology enrichment resources available on PubMed . A total of 686 genes from cluster analysis were assigned GO terms . After filtering using the significant criterion of p,0.05, we selected 265 genes that fell in the range of key functional classifications. A Comparison of the comprehensive GO analysis and cluster GO analysis, suggests that immune response , cellular component organization, response to stimuli, cellular processes and signaling processes are the most prominent cellular functions during recovery. An interaction network of significant GO terms was assembled into a GO map to depict the relationship among prominent functional categories . Based on the GO map, we can directly and systematically find the subordinate relationship between GO terms . There are 12 subnetworks clustered into four functions: 1) immune : T/B cell activation and immune response; 2) response to stimuli: response to stress and defense response to bacterium; 3) cellular process: cell differentiation, proliferation, migration, adhesion and cycle; 4) cell signaling: protein amino acid phosphorylation, negative regulation of transcription factor activity and cyclic nucleotide catabolic process. To further narrow down the gene data set for genes that regulate bone marrow recovery, 265 genes selected in GO analysis were analyzed by gene co-expression network with k-core algorithm . Gene co-expression networks were built according to the normalized signal intensity of differentially expressed genes. The network shown in Figure 4 reflects the correlations between genes. Each node describes a given gene, and the relationship between a pair of genes is represented with a line segment.