Other pieces poll others’ status: koji scrapes pkgdb for package-owner relationships and email aliases.
This dense coupling of services makes changing, adding, or replacing services more complicated: commits to one project require a spidering of code changes to all the others.
For a given genome, the search of synteny groups is performed, firstly with genome data stored into Pk GDB, and secondly with the proteomes of other complete genomes of the Ref Seq section in Gen Bank (whose data are then integrated into Pk GDB in their original version). In the case of the annotation of a newly sequenced genome, our databases are used to manage sequence data being on finishing stage of the sequencing process that is still difficult, especially for GC rich genomes or containing many repetitions.
jkeating’s talk on messaging in the Fedora Instructure proposed the adoption of a unified message bus to reduce the complexity of multiple interdependent services.
Instead of a service interfacing with its dependencies’ implementations, it could subscribe to a topic whenever an account’s email address changes.
A second set of relational tables in Pk GDB can then store the results of the AMIGene method whose set of CDS predicted with model calculations is compared to annotated genes in banks: this way we identify non-annotated genes (unique annotation in AMIGene on which Blast comparisons are automatically generated), and genes that may correspond to erroneous annotations (annotation of sequences unique to banks).
In particular, groups of synteny with other genomes are then sought, taking into account: Fragments of genes identified in the chromosome (which are however no more functional in the cell).