@article{Nagashima01062003,
author = {Nagashima, Takeshi and Silva, Diego G. and Petrovsky, Nikolai and Socha, Luis A. and Suzuki, Harukazu and Saito, Rintaro and Kasukawa, Takeya and Kurochkin, Igor V. and Konagaya, Akihiko and Schönbach, Christian}, 
title = {Inferring Higher Functional Information for RIKEN Mouse Full-Length cDNA Clones With FACTS},
volume = {13}, 
number = {6b}, 
pages = {1520-1533}, 
year = {2003}, 
doi = {10.1101/gr.1019903}, 
abstract ={FACTS (Functional Association/Annotation of cDNA Clones from Text/Sequence  Sources) is a semiautomated knowledge discovery and annotation system that  integrates molecular function information derived from sequence analysis  results (sequence inferred) with functional information extracted from text.  Text-inferred information was extracted from keyword-based retrievals of  MEDLINE abstracts and by matching of gene or protein names to OMIM, BIND, and  DIP database entries. Using FACTS, we found that 47.5% of the 60,770 RIKEN  mouse cDNA FANTOM2 clone annotations were informative for text searches.  MEDLINE queries yielded molecular interaction-containing sentences for 23.1%  of the clones. When disease MeSH and GO terms were matched with retrieved  abstracts, 22.7% of clones were associated with potential diseases, and 32.5%  with GO identifiers. A significant number (23.5%) of disease MeSH-associated  clones were also found to have a hereditary disease association (OMIM  Morbidmap). Inferred neoplastic and nervous system disease represented 49.6%  and 36.0% of disease MeSH-associated clones, respectively. A comparison of  sequence-based GO assignments with informative text-based GO assignments  revealed that for 78.2% of clones, identical GO assignments were provided for  that clone by either method, whereas for 21.8% of clones, the assignments  differed. In contrast, for OMIM assignments, only 28.5% of clones had  identical sequence-based and text-based OMIM assignments. Sequence, sentence,  and term-based functional associations are included in the FACTS database  (http://facts.gsc.riken.go.jp/),  which permits results to be annotated and explored through web-accessible  keyword and sequence search interfaces. The FACTS database will be a critical  tool for investigating the functional complexity of the mouse transcriptome,  cDNA-inferred interactome (molecular interactions), and pathome  (pathologies).}, 
URL = {http://genome.cshlp.org/content/13/6b/1520.abstract}, 
eprint = {http://genome.cshlp.org/content/13/6b/1520.full.pdf+html}, 
journal = {Genome Research} 
}