Bioconductor maintains both platform specific and general annotation data packages that can be used independantly or jointly. Contents of platform specific annotation data packages are show below using the annotation package for AffyMetrix HG U95Av2 (hgu95av2) chip as an example. Contents of general annotation pacakges (homology, organismCHRLOC, organismLLMappings, GO, and KEGG, where organism can be human, mouse, or rat) are listed individually.
Annotation data packages can be built using Bioconductor's AnnBuilder library. The Vignettes sections contains instructions on how to build data packages using AnnBuilder.
Bioconductor's Annotation Packages are built based on data retrieved from the following public data repositories:
The following sources are used to derive unified mappings between probe ids and LocusLink ids:
Construction
For each set of target ids (probe) that needs to be annotated, unified mappings between the target ids and LocusLink ids are sought by combining the available maps provided by dChip, Gene Ontology, Unigene annotation for Affy chips and those that are derived from data provided by LocusLink and Unigene. The unified mappings of target ids to LocusLink ids are then used as the points of linkage to obtain annotation data from the sources listed above.
An annotation package is a regular R package that contains data, help files, and R functions and can be installed/loaded the same way as an R package. Annotation data are contained as binary hash table like files in the data subdirectory.Each file has either probe ids or (in the case of reverse mapping) the values of an annotation element as keys and the corresponding values of an annotation element or (in the case of reverse mapping) probe ids as values.Detailed description of each binary file and its usage can be found in the help file.