An improved systematic approach to predicting transcription factor target genes using support vector machine.

Area of Science:

• Computational Biology
• Bioinformatics
• Genomics

Background:

• Understanding gene regulatory networks is crucial for biological insights.
• Experimental methods for identifying transcription factor target genes (TFTGs) are time-consuming and laborious.
• Existing computational approaches often have limited performance and neglect dataset structural properties.

Purpose of the Study:

• To develop a refined, systematic computational approach for predicting TFTGs.
• To improve the accuracy and efficiency of TFTG prediction compared to existing methods.
• To address limitations of current algorithms, particularly regarding dataset structure.

Main Methods:

• Utilized a novel reverse-complementary distance-sensitive n-gram profile algorithm.
• Converted upstream DNA sub-sequences into high-dimensional vector data points.
• Employed a support vector machine classifier for the prediction task, treating it as a classification problem.

Main Results:

• The proposed approach demonstrated significant performance improvement over other computational methods.
• Validated using 10-fold cross-validation, showing a higher area under the curve (AUC) on the receiver operating characteristic (ROC) curve.
• Precision-recall and cost curves also yielded highly satisfactory results, addressing dataset skewness.

Conclusions:

• The refined computational approach offers a more effective method for predicting TFTGs.
• This advancement contributes to a better understanding of gene regulatory networks.
• The method shows promise for overcoming limitations of previous computational strategies in bioinformatics.