Review of PubGene
Date Reviewed: May 20,2001
Description: This is a database for querying the co-occurrence of human gene symbols in literature. It discovers the relationship or network of the interest gene between other genes. The database can be used to: 1) Analyze gene expression data with literature network information, 2) Browse literature neighbors of a given gene, 3) Search literature articles for a set of genes, 4) Search ontology terms related to a given gene, 5) Search MeSH terms found with a set of genes, 6) Search for official nomenclature
Strengths: This will be a good resource to find the network of a gene. It will save researchers a lot of time to search the literature and find the possible relationship of the interest genes. The database can give you a general idea what are the related genes of your target gene.
Weaknesses: The database is still incomplete. The main reason may come from the method that is used for mining the co-occurrence of gene symbols in literatures. Gene occurrences are found by searching for gene symbols and short gene names, such as single word or names of the form ‘family variant’. The trouble is that if the gene has long name consisting of several words, such as mannose 6-phosphate receptor, the search tool may miss this. This will lead to false negative. The other problem may be that one name may represent multiple genes, which will lead to false positive. My conclusion is that you have to prove the result from the database by searching literature by yourself.
Comprehensiveness: The database covers 13712 human genes extracted from the titles and abstracts over 10 million article records from the medline citation database.
Ease of Use: good
Similar or Related Sites:Database of interacting Proteins(DIP) and the online Mendelian Inheritance in man database(OMIM)
Overall Evaluation: ***
Reviewer: Jin Li