Figure 4

The basics of community annotation and semantic support. Once Knowlets have been created from authoritative sources and the indexed literature, a regular re-computing of the concept space with all changed semantic associations is performed. In case new co-occurrences, stated facts or significant associations emerge from the computational process, all experts that have expressed an interest in that part of the concept space will be alerted. Pre-constructed Knowlets for over one million authors have been created who currently have publications in PubMed. When they comment in the Wiki, their contributions will automatically be indexed and processed, forming an additional source for Knowlet enrichment alongside the classic literature and databases. UniProtKB/Swiss-Prot, GO Annotation, IntAct and UMLS have indicated that they wish to use the system as a source for accelerated annotation in their respective information resources.