Gene regulation research concerns the regulatory relationship between transcription factors (TFs) and their target genes (TGenes). Due to the rapid acceleration of biological research, it is impractical for biologists to read all of the relevant literature and manually extract all of the information about the regulatory relationships between a TF and its TGenes. This paper proposes a method utilizing negative and positive textual patterns to extract regulatory information regarding certain TF-TGene pairs, which provides insightful information to biologists and saves them time from excessive literature reading. We hypothesized that the negative patterns could be used for filtering and that the system would mainly rely on the positive patterns to mine the regulatory TF-TGene relationships from the text. We also examined whether WordNet could be utilized to improve the pattern recognition performance. The results show that the negative pattern should be used for initial filtering, and then the positive patterns can extract information related to gene regulation. Moreover, WordNet seems to have little effect on the performance when extracting gene regulations.
All Science Journal Classification (ASJC) codes
- Computer Science Applications
- Health Informatics