Project Description

This site will provide access to an apoptosis database containing extant protein sequences, structures, models, nomenclature and other information pertaining to apoptosis. It is intended as a bioinformatics resource for the apoptosis research community.

This resource is provided via the Burnham Institute and funded by the National Science Foundation. It is being developed in the labs of Profs. John C. Reed, Philip E. Bourne & Adam Godzik by Kutbuddin S. Doctor at The Burnham Institute.

Your feedback is greatly appreciated.

As of April 2002 the site has the following functionality:

• Thirteen domains involved in apoptosis being maintained.
• The NCBI non-redundant database is searched for homologs for each domain. Results of several searches are gathered, further annotated, and categorized. Listing from this database are available.
• Proteins are brought into the database based on homology for a domain via profile based sequence homology methods. Principally by Psi-BLAST, assisted by Saturated Blast (developed by A. Godzik, W. Li) and often confirmed by Pfam and Pfscan
• Proteins, after being brought in, are annotated via homology to families of full length proteins. A family is founded by one or several full length proteins sharing similar set of domains each of which are distinctly clustered (by phylogeny) and are either known orthologs or clearly closest homologs. We use the Family Pairwise Search (FPS) algorithm by W. Grundy and T. Bailey to cluster all other protein with respect to these families only over their apoptotic domains. Thus, membership to a family requires having closest homology to that family over an apoptosis related domain.
• This differs from domain Architecture tools (like DART at NCBI, SMART by P. Bork, or InterPro) which consider give only one model for a given domain. In this resource we break up the single model of a InterPro or Pfam domain into a set of models that represent functional and phlogenetic groupings.
• The database of sequences maintained by apoptosis-db.org can be searched using BLAST.
• The annotation in the database can be searched.
• Each protein in the database also has a graphical report of the domains (including non-apoptosis domains) found by, thus far, three tools. The domains can overlap (example)
• If strong enough homology to a structure is found via FFAS (Godzik group), then a structural model is created and maintained via automated modelling by the Godzik group Modeller server.
• A map of the web site and connectivity
• Current statistics on Proteins and Domains within this database