Associations with GLEANR gene prediction models (which comprise Release 2.1) were determined for previously annotated genes in D. pseudoobscura (Release 2.0; FBrf0184012). This analysis consisted of both automated and manual steps. The initial automated assessment identified existing annotations and GLEANR predictions for which there were reciprocal 1:1 overlaps (N=9243). For these cases, the original "GA" annotation identifer and FlyBase "FBgn" gene identifier have been retained, but the Release 2.0 gene model has been replaced by the GLEANR gene prediction model.

The remaining ambiguous cases were assessed manually. In cases of multiple or complex overlaps, new primary identifiers (GA annotation numbers and FBgn gene ID's) were assigned; the existing identifiers were retained as synonyms (N=690). Comments were added noting cases of annotation merges, splits, or complex merge+splits. A portion of the Release 2.0 D. pseudoobscura annotations consisted of defined genomic extents (regions containing sequences orthologous to a D. melanogaster gene), without a transcript model. These have been marked as "withdrawn" (N=2271). A smaller number of Release 2.0 annotations which included transcript models, but did not overlap GLEANR gene predictions, were also marked as withdrawn (N=105); after reassessment some of these may be reinstated in a future release.

Associations with GLEANR gene prediction models were also determined for previously described genes for which there were no annotations, but for which nucleotide or protein sequence data are available (GenBank accessions or data from personal communications to FlyBase; N=44).