Meta, Meta(N) and Cyber servers

摘要

The recent CASP (2002) meeting highlighted the trends emerging in the protein structure community, revealing a hidden conflict that may be of interest to the more general biology and bioinformatics communities. Bioinformatics offers the biologist a large number of computational methods implemented in the form of web servers. In particular, almost all the protein-structure prediction methods are available on the web. These facilities have also been exploited by a second generation of Meta servers that refine their predictions by combining results from various basic servers. The range of developments includes: simple majority-vote schemas, consensus predictions and elaborate filters using external information (e.g. analysis of multiple sequence alignment, or functional sequence features). Furthermore other servers mix and recombine the information provided by the Meta servers, in what can be considered an exponential tendency (MetaN servers).