Figure 6

Flow of information during advanced data exploration using macros. A user wants to find all scientific articles related to protein sequences that are the translation products of nucleotide sequences corresponding to HNF1-alpha (hepatocyte nuclear factor 1-alpha). A user defines a macro by specifying: 1) initial query to be executed against the first of the chosen databases (e.g. hnf1a[Title] AND human[Organism]), 2) all databases that he/she wants to explore in order to find related information he/she is interested in, and 3) the order of database access (Nucleotide, Protein, PubMed). On the basis of the information provided by a user, search GenBank translates the macro into the choreography of calls of appropriate tools available through the NCBI Web Services. An appropriate pipeline of successive calls is created, including ESearch, ELink and ESummary tools. The ESearch component runs the query provided by the user against the Nucleotide database. Results of the query execution are passed to the first ELink, which searches related records in the Protein database, and again, the returned records (UID s) parameterize the call of the second ELink component, which finds related articles in the PubMed database. Finally, the summary of articles is returned by the ESummary component in the form of the XML document. This document is sent back to the search GenBank, which presents it to the user in a friendly form. After a short review of returned records, the user decides to see full details of one of them (e.g. UID 22517943). He/she interactively triggers a call of EFetch tool, which retrieves data from the PubMed database for a passed record identifier UID and presents details of the chosen record on the search GenBank web page.