| Summary: | PubServer, available at http://pubserver.burnham.org/, is a tool to automatically
collect, filter and analyze publications associated with groups of homologous
proteins. Protein entries in databases such as Entrez Protein database at NCBI
contain information about publications associated with a given protein. The
scope of these publications varies a lot: they include studies focused on
biochemical functions of individual proteins, but also reports from genome
sequencing projects that introduce tens of thousands of proteins. Collecting and
analyzing publications related to sets of homologous proteins help in functional
annotation of novel protein families and in improving annotations of
well-studied protein families or individual genes. However, performing such
collection and analysis manually is a tedious and time-consuming process.
PubServer automatically collects identifiers of homologous proteins using
PSI-Blast, retrieves literature references from corresponding database entries
and filters out publications unlikely to contain useful information about
individual proteins. It also prepares simple vocabulary statistics from titles,
abstracts and MeSH terms to identify the most frequently occurring keywords,
which may help to quickly identify common themes in these publications. The
filtering criteria applied to collected publications are user-adjustable. The
results of the server are presented as an interactive page that allows
re-filtering and different presentations of the output.
|
|---|