The HAMAP project is intended to annotate the microbial proteomes presently available (see the "Summary of microbial proteomes". Each proteome contains thousands of proteins which can be partly grouped into functional families implied in the various biological processes that govern a microbial organism (namely its metabolism).
Few microbial metabolic pathway have already been defined in terms of HAMAP functional families and number of others are under study.
Checking the consistency of the annotation at the organism level therefore means dealing with the metabolic pathways of the microbial organism.
For this purpose, and in the context of the Sibelius association between Helix and the Swiss-Prot group in Geneva, we have developed a tool named HERBS, for "HAMAP Expert Rule Based System". This application is designed in order to help the users:
increasing the knowledge on metabolic pathways by comparing them in different microbial organisms.
checking the consistency of the annotation of all described metabolic pathways for one particular microbial organism.
This application is available into two versions: a console version and a graphical one. Both versions provide results in the form of texts and graphs with special emphasis on warning the users of missing, unexpected, ambiguous and also normal proteins:
The missing proteins are expected to be observed in the studied organism, but turn out to be absent.
The unexpected proteins are not expected in the studied organism, but turn out to be present.
The ambiguous proteins are both present and required since they are implied in an expected metabolic pathway in the organism, but are neverttheless unexpected since they are involved in a biological process that is not expected in the organism.
The normal proteins are those which are both required and present.
The graphical HERBS version also offers the users the ability:
to get up to date data from the database he selected
to visualise and edit graphically metabolic pathways in a simple way
to select and test some data inference criteria
It is made up of:
an inference engine based on Jess and a knowledge base, i.e. facts and rules;
a graphical friendly user interface.
A first version of HERBS will be soon open source on this site.
Besides the improvement of the HERBS application, the progressive extension of the knowledge base will provide larger sets of data and will subsequently contribute to improve the annotation of the microbial proteomes.