Single GP Process (Archived)
The Problem
A conserved metabolic pathway appears in several different species; we can represent the steps using the 'has part' relationship. In one species, the entire pathway is catalyzed by a single gene product; in another, several GPs work together to perform the pathway. In the former situation, where the single gene product performs all the reactions in the pathway, it would be useful to be able to annotate to the pathway term and have the annotations automatically propagate to the "has part" children.
The Possible Solution
The AP ("All Parts") modifier, which, when applied to an annotation, indicates that a GP is responsible for the entire process (all parts of the process), and that annotations can be propagated down to all "has part" children. No propagation occurs in the absence of the qualifier.
Example
Fatty acid biosynthesis by fatty acid synthase enzymes. The basic pathway is a cycle of the following enzyme activities:
elongation reaction --> ketoreductase --> dehydratase --> enol reductase --> (back to elongation)
From the Fatty Acid Synthase article on Wikipedia:
There are two principal classes of fatty acid synthases.
Type I systems utilise a single large, multifunctional polypeptide and are common to both mammals and fungi (although the structural arrangement of fungal and mammalian synthases differ). A Type I fatty acid synthase system is also found in the CMN group of bacteria (corynebacteria, mycobacteria, and nocardia). In these bacteria, the FAS I system produces palmititic acid, and cooperates with the FAS II system to produce a greater diversity of lipid products.
Type II is found in archaeabacterial and eubacterial, and is characterized by the use of discrete, monofunctional enzymes for fatty acid synthesis. Inhibitors of this pathway (FASII) are being investigated as possible antibiotics.
The mechanism of FAS I and FAS II elongation and reduction is the same, as the domains of the FAS II enzymes are largely homologous to their domain counterparts in FAS I multienzyme polypeptides. However, the differences in the organization of the enzymes - integrated in FAS I, discrete in FAS II - gives rise to many important biochemical differences.
(my emphasis)
In this case we would have an ontology structure something like this:
fatty acid biosynthesis -has part- elongase activity -has part- ketoreductase activity -has part- dehydratase activity -has part- enol reductase activity
To annotate type II fatty acid synthases, you would have a number of gene products, each of which would have a different function annotation; all would also be annotated to fatty acid biosynthesis. No propagation of annotations from the fatty acid biosynthesis term.
To annotate type I fatty acid synthases, where a single GP performs ALL the functions, you would annotate to fatty acid biosynthesis with the All Parts modifier. This indicates that the annotations can automatically be propagated to all the "has part" children of the term.