PNAS: Protein ranking: from local to global structure in the protein similarity network [clip]

A paper of interest:

In Weston et al., Noble and colleagues develop an algorithm for
comparing protein sequences that has inspired from this similarity
measures used by the Google search engine. They use a diffusion
distance type of approach that takes into account the overall
connectivity of similarity relationships. This paper represents a
great synthesis of the way a computer scientist can apply general
methods developed outside of biology into the biological context.

--
Weston J, Elisseeff A, Zhou D, Leslie CS, Noble WS. Protein ranking:
from local to global structure in the protein similarity network.
Proc Natl Acad Sci U S A. 2004 Apr 27;101(17):6559-63. Epub 2004 Apr
15.