Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Identificación de Nombres de Genes en la Literatura Biomédica

940 views

Published on

Carmen Galvez, Félix Moya-Anegón
Scimago Research Group, University of Granada, Faculty of Library and Information Science, Granada, SPAIN

Published in: Education, Technology
  • Be the first to comment

  • Be the first to like this

Identificación de Nombres de Genes en la Literatura Biomédica

  1. 1. Gene Name Identification in Biomedical Literature Carmen Galvez & Félix Moya-Anegón {cgalvez, felix}@ugr.es SCImago Research Group, University of Granada SPAIN
  2. 2. Motivation <ul><li>Retrieving relevant documents for genes is important </li></ul><ul><ul><li>Information Extraction (Shatkay et. al. 2001) </li></ul></ul><ul><ul><li>Text mining (Srinivasan et. al. 2004) </li></ul></ul><ul><li>Retrieval affected by ambiguity in gene nomenclature </li></ul><ul><ul><li>Homonymy </li></ul></ul><ul><ul><ul><li>AFP - official symbol for encoding alpha-fetoprotein, also an alias for another gene of the TRIM family. </li></ul></ul></ul><ul><ul><li>Synonymy </li></ul></ul><ul><ul><ul><li>AGRP: ART, ASIP2 </li></ul></ul></ul><ul><ul><li>Common English words as gene terms </li></ul></ul><ul><ul><ul><li>BAD, AM, TRAP, ACHE, BAR, ANT, APEX </li></ul></ul></ul>
  3. 3. Introduction <ul><li>Gene term ambiguity is a well known problem </li></ul><ul><li>Various approaches have been used </li></ul><ul><ul><li>Machine Learning (Hatzivassiloglou et. al. 2001) </li></ul></ul><ul><ul><li>Hidden Markov Model (Morgan et. al. 2003) </li></ul></ul><ul><ul><li>Complex rule-based systems (Tanabe et. al. 2002) </li></ul></ul>
  4. 4. Method <ul><li>Step 1 . Collection of gene terms using a Biological Database. </li></ul><ul><li>Step 2 . Encoding of gene-naming terms in a Binary Matrix. </li></ul><ul><li>Step 3 . Design of a Parametrized Finite-State Graph (P-FSG) that formalizes the gene term variants encoded in the Binary Matrix. </li></ul>
  5. 5. Method – Step 1 <ul><li>Step 1. Collection of gene terms using a Biological Database (FlyBase) </li></ul>LIST OF GENE SYNONYMS ID NUMBER
  6. 6. Method – Step 2 <ul><li>Step 2. Encoding of gene-naming terms in a Binary Matrix </li></ul>
  7. 7. Method – Step 3 <ul><li>Step 3. Design of a Parametrized Finite-State Graph (P-FSG) </li></ul>
  8. 8. Unified Gene Symbols <ul><li>Acf=> FBgn0027620 </li></ul><ul><li>ACF=> FBgn0027620 </li></ul><ul><li>ATP=> FBgn0027620 </li></ul><ul><li>CAF=> FBgn0027620 </li></ul><ul><li>CHRAC=> FBgn0027620 </li></ul><ul><li>dACF=> FBgn0027620 </li></ul><ul><li>dCHRAC=> FBgn0027620 </li></ul><ul><li>acf 1=> FBgn0027620 </li></ul><ul><li>ACF 1=> FBgn0027620 </li></ul><ul><li>Acf 1=> FBgn0027620 </li></ul><ul><li>Acf - 1=> FBgn0027620 </li></ul><ul><li>Chromatin Accessibility Complex=> FBgn0027620 </li></ul><ul><li>CG 1 9 6 6=> FBgn0027620 </li></ul><ul><li>CHRAC - 1 7 5=> FBgn0027620 </li></ul><ul><li>p 1 7 0 / p 1 8 5=>FBgn0027620 </li></ul>
  9. 9. Results – Detected Gene Symbols in Medline <ul><li>Binding of Acf1 to DNA involves a WAC motif and is important for ACF mediated chromatin assembly.Fyodorov,-D-V; Kadonaga,-J-TMol-Cell-Biol. 2002 Sep; 22(18): 6344-53 </li></ul><ul><li> </li></ul><ul><li> ACF is a chromatin-remodeling complex that catalyzes the ATP -dependent assembly of periodic nucleosome arrays. This reaction utilizes the energy of ATP hydrolysis by ISWI, the smaller of the two subunits of ACF . Acf1 , the large subunit of ACF , is essential for the full activity of the complex. We performed a systematic mutational analysis of Acf1 to elucidate the functions of specific subregions of the protein. These studies revealed DNA- and ISWI-binding regions that are important for the chromatin assembly and ATPase activities of ACF . The DNA-binding region of Acf1 includes a WAC motif, which is necessary for the efficient binding of ACF complex to DNA. The interaction of Acf1 with ISWI requires a DDT domain, which has been found in a variety of transcription and chromatin-remodeling factors. Chromatin assembly by ACF is also impaired upon mutation of an acidic region in Acf1 , which may interact with histones during the deposition process. Lastly, we observed modest chromatin assembly defects on mutation of other conserved sequence motifs. Thus, Acf1 facilitates chromatin assembly via an N-terminal DNA-binding region with a WAC motif, a central ISWI-binding segment with a DDT domain, and a C-terminal region with an acidic stretch, a WAKZ motif, PHD fingers, and bromodomain </li></ul>
  10. 10. Results – Unified Gene Symbols in Medline <ul><li>Binding les Acf1 to DNA involves a WAC motif and les important for ACF mediated chromatin assembly.Fyodorov,-D-V; Kadonaga,-J-TMol-Cell-Biol. 2002 Sep; 22(18): 6344-53 </li></ul><ul><li> FBgn0027620 is a chromatin-remodeling complex that catalyzes the FBgn0027620 -dependent assembly of periodic nucleosome arrays. This reaction utilizes the energy of FBgn0027620 hydrolysis by ISWI, the smaller of the two subunits of FBgn0027620 . FBgn0027620 , the large subunit of FBgn0027620 , is essential for the full activity of the complex. We performed a systematic mutational analysis of FBgn0027620 to elucidate the functions of specific subregions of the protein. These studies revealed DNA- and ISWI-binding regions that are important for the chromatin assembly and ATPase activities of FBgn0027620 . The DNA-binding region of FBgn0027620 includes a WAC motif, which is necessary for the efficient binding of FBgn0027620 complex to DNA. The interaction of FBgn0027620 with ISWI requires a DDT domain, which has been found in a variety of transcription and chromatin-remodeling factors. Chromatin assembly by FBgn0027620 is also impaired upon mutation of an acidic region in FBgn0027620 , which may interact with histones during the deposition process. Lastly, we observed modest chromatin assembly defects on mutation of other conserved sequence motifs. Thus, FBgn0027620 facilitates chromatin assembly via an N-terminal DNA-binding region with a WAC motif, a central ISWI-binding segment with a DDT domain, and a C-terminal region with an acidic stretch, a WAKZ motif, PHD fingers, and bromodomain. </li></ul>
  11. 11. Conclusion <ul><li>The ambiguity produced by homonyms and English words has a detrimental effect on precision. </li></ul><ul><li>Disambiguation methods are highly complex, and generally distinguish the context where the gene names appear, as: </li></ul><ul><ul><li>Hidden Markov Model (HMM)-based taggers that consider the probability conditioned by the context. </li></ul></ul><ul><ul><li>Methods based on contextual similarity relations through co-occurrence analysis. </li></ul></ul>
  12. 12. Carmen Galvez & Félix Moya-Anegón {cgalvez, felix}@ugr.es SCImago Research Group, University of Granada SPAIN

×