Limits...
Detection of gene interactions based on syntactic relations.

Kim MY - J. Biomed. Biotechnol. (2008)

Bottom Line: While the first phase contributes to improve recall, the second and third phases contribute to improve precision.In the experimental results using ICML 05 Workshop on Learning Language in Logic (LLL05) data, our proposed method gave an F-measure of 67.2% for the test data, significantly outperforming previous methods.We also describe the contribution of each phase to the performance.

View Article: PubMed Central - PubMed

Affiliation: School of Computer Science and Engineering, Sungshin Women's University, Seoul 136-742, Korea. miykim@sungshin.ac.kr

ABSTRACT
Interactions between proteins and genes are considered essential in the description of biomolecular phenomena, and networks of interactions are applied in a system's biology approach. Recently, many studies have sought to extract information from biomolecular text using natural language processing technology. Previous studies have asserted that linguistic information is useful for improving the detection of gene interactions. In particular, syntactic relations among linguistic information are good for detecting gene interactions. However, previous systems give a reasonably good precision but poor recall. To improve recall without sacrificing precision, this paper proposes a three-phase method for detecting gene interactions based on syntactic relations. In the first phase, we retrieve syntactic encapsulation categories for each candidate agent and target. In the second phase, we construct a verb list that indicates the nature of the interaction between pairs of genes. In the last phase, we determine direction rules to detect which of two genes is the agent or target. Even without biomolecular knowledge, our method performs reasonably well using a small training dataset. While the first phase contributes to improve recall, the second and third phases contribute to improve precision. In the experimental results using ICML 05 Workshop on Learning Language in Logic (LLL05) data, our proposed method gave an F-measure of 67.2% for the test data, significantly outperforming previous methods. We also describe the contribution of each phase to the performance.

Show MeSH
Reverse syntactic path of Figure 1 for a negative rule.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2277490&req=5

fig5: Reverse syntactic path of Figure 1 for a negative rule.

Mentions: We obtain twotypes of rule set. One is a positive rule set obtained by learning thedirection from an agent to its target. The other is a negative rule setobtained by learning the direction from a target to its agent in reverse order.Figure 2 shows the reverse syntactic path from a target to its agent of thesentence in Figure 1. The positive and negative rules for the sentence inFigure 1 are shown in Table 1. From the positive and negative rule sets, weconstruct direction rules according to the following subsections.


Detection of gene interactions based on syntactic relations.

Kim MY - J. Biomed. Biotechnol. (2008)

Reverse syntactic path of Figure 1 for a negative rule.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2277490&req=5

fig5: Reverse syntactic path of Figure 1 for a negative rule.
Mentions: We obtain twotypes of rule set. One is a positive rule set obtained by learning thedirection from an agent to its target. The other is a negative rule setobtained by learning the direction from a target to its agent in reverse order.Figure 2 shows the reverse syntactic path from a target to its agent of thesentence in Figure 1. The positive and negative rules for the sentence inFigure 1 are shown in Table 1. From the positive and negative rule sets, weconstruct direction rules according to the following subsections.

Bottom Line: While the first phase contributes to improve recall, the second and third phases contribute to improve precision.In the experimental results using ICML 05 Workshop on Learning Language in Logic (LLL05) data, our proposed method gave an F-measure of 67.2% for the test data, significantly outperforming previous methods.We also describe the contribution of each phase to the performance.

View Article: PubMed Central - PubMed

Affiliation: School of Computer Science and Engineering, Sungshin Women's University, Seoul 136-742, Korea. miykim@sungshin.ac.kr

ABSTRACT
Interactions between proteins and genes are considered essential in the description of biomolecular phenomena, and networks of interactions are applied in a system's biology approach. Recently, many studies have sought to extract information from biomolecular text using natural language processing technology. Previous studies have asserted that linguistic information is useful for improving the detection of gene interactions. In particular, syntactic relations among linguistic information are good for detecting gene interactions. However, previous systems give a reasonably good precision but poor recall. To improve recall without sacrificing precision, this paper proposes a three-phase method for detecting gene interactions based on syntactic relations. In the first phase, we retrieve syntactic encapsulation categories for each candidate agent and target. In the second phase, we construct a verb list that indicates the nature of the interaction between pairs of genes. In the last phase, we determine direction rules to detect which of two genes is the agent or target. Even without biomolecular knowledge, our method performs reasonably well using a small training dataset. While the first phase contributes to improve recall, the second and third phases contribute to improve precision. In the experimental results using ICML 05 Workshop on Learning Language in Logic (LLL05) data, our proposed method gave an F-measure of 67.2% for the test data, significantly outperforming previous methods. We also describe the contribution of each phase to the performance.

Show MeSH