7. Gene ontology database: structure and validation of the data...

• Look at the definition of ontology in wikipedia
• Go to the Gene ontology consortium (AmiGO)
• Compare the GO terms associated with mouse and human erythropoeitin
• Go to the UniProtKB/Swiss-Prot entry for human erythropoietin: look at the associated GO terms How many terms have been 'manually' attributed to the gene (have a look to the GO evidence tag )?

• Look at the definition of ontology in wikipedia
• Go to the Gene ontology consortium (AmiGO)
• Compare the GO terms associated with mouse and human erythropoeitin :
  
  
  
• Go to the UniProtKB/Swiss-Prot entry for human erythropoietin: look at the associated GO terms. How many terms have been 'manually' attributed to the gene (have a look to the GO evidence tag )?

Remark: the GO database is of good quality. The problem is the usage of these GO terms: they are attributed automatically to a given gene, according to the results of protein sequence analysis (=prediction), i.e. domains (InterPro results), subcellular location etc…