BDGP in situ annotation data was incorporated in two stages during the production of FlyBase releases FB2012_05 and FB2012_06. BDGP expression annotations and cDNA lists were downloaded from the BDGP in situ webpage was downloaded August 1, 2012. Integration of the BDGP data with FlyBase data required two transformations. First, differences in the anatomy controlled vocabularies were reconciled. Second, BDGP cDNA-gene assignments were re-evaluated to account for changes to gene models since the beginning of the BDGP in situ project.

A total of 325 unique anatomy terms were used by the BDGP in 81,952 annotations. 148 BDGP terms were exact matches to current FlyBase anatomy terms (23,575 annotations). 115 were nearly identical matches or easily converted into FlyBase syntax (33,725 annotations). 46 BDGP terms differed in the use of terms anlage in statu nascendi, anlage and primordium (2,086 annotations). 10 terms were imperfectly mapped to the most closely related FlyBase anatomy term (1,417 annotations). In such cases where the BDGP and FlyBase terms differ significantly, the originally reported BDGP term is indicated in the comments attached to the FlyBase expression statement. Six terms were not curated, including 20,915 annotations of "no staining", 221 annotations involving two internal BDGP flags, and three other infrequent terms judged to be ambiguous (13 annotations). See this file, listed below, for details: FlyBase_Analysis_BDGP_insitu_annotation_term_mapping.120817.xlsx

A total of 7,775 cDNA/ESTs were used by the BDGP for in situ expression analysis. Because FlyBase gene models have changed substantially since the BDGP project first began, these cDNA/ESTs were re-mapped to Dmel 5.46 (FlyBase release FB2012_04)gene models: 7,539 of 7,775 cDNA/ESTs were successfully mapped. In some cases, a cDNA/EST was associated with multiple genes: cDNA/ESTs representing dicistronic gene models were associated with both genes of the dicistronic module; cDNA/ESTs from genes with highly related paralogs in the genome were linked to all genes with 99% identity to the cDNA/EST. cDNAs could not be mapped to genes if they were artifactual chimeric cDNAs composed of fragments from two or more genes or transposons, or, if they did not map to an annotated transcript. See this file, listed below, for details: FlyBase_Analysis_BDGP_insitu_cDNA-to-gene_mapping.120817.xlsx