Information Technology Reference
In-Depth Information
Fig. 5. Class of extracted interaction information at whole HLA and whole disease:
the number of parenthesis is the count of sentences
the total number of extracted HLA-disease interactions and categorized num-
bers. Extracted interaction information is displayed in Fig.6. We offer level 3
and level 2 data. Level 2 data is sometimes considered as not usable data but
their statistic knowledge is important for further research. Additionally, the num-
ber of HLA-disease interactions according to geographic location demonstrates
how many research studies have been carried out in those geographic locations.
Furthermore, we offer some usable tools like view parsetree for all sentences,
view abstracts which contain selected sentences and more.
Since HLA entity name were listed in various transformation, it is quite di-
cult to recognize HLA entity by using the regular expression. For example, the
sentence 'Haplotype-specific patterns were defined with DNA from DR1, 2, 3,
4, 7, 8, 11, 12, and 13 homozygous typing cells, with restriction enzymes Eco
RI, Bgl I, and Pvu II.' have 9 DR antigen. But we identified only DR1. This
problems appeared to the sentence 'DQB1*03/*04/*05/*06, DQA1*0301 (Ls,
Stage II) were less common than in the controls.'.
In spite of the fact that MeSH is a rich resource in biomedical keyword search,
it does not contain much information related to disease entity. For this reason
our system was low disease recognition. But we should expect more confident
recognition when disease recognition is developed by using the UMLS, a resource
of the language of biomedicine and health and MeSH is used only search in result.
Search WWH ::




Custom Search