Related Experiment Videos
RSIR: regularized sliced inverse regression for motif discovery
Wenxuan Zhong1, Peng Zeng, Ping Ma
1Department of Statistics, Harvard University, Cambridge, MA 02138, USA.
Bioinformatics (Oxford, England)
|September 17, 2005
Summary
We introduce regularized sliced inverse regression (RSIR), a novel computational method for identifying transcription factor binding motifs (TFBMs). RSIR enhances motif detection by integrating gene expression and promoter sequence data, offering improved accuracy and stability.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Identifying transcription factor binding motifs (TFBMs) is essential for understanding gene regulation.
- Current methods face challenges with high-dimensional and collinear data.
Purpose of the Study:
- To propose a novel computational procedure, regularized sliced inverse regression (RSIR), for identifying TFBMs.
- To leverage both gene expression and promoter sequence data for improved motif detection.
Main Methods:
- The study introduces regularized sliced inverse regression (RSIR).
- RSIR combines gene expression data with promoter sequence information.
- The method is designed for computational efficiency and stability with high-dimensional data.
Main Results:
- RSIR demonstrates a lower false positive rate compared to SIR and stepwise regression on simulated data.
- The method shows excellent performance when applied to yeast amino acid starvation and cell cycle datasets.
Conclusions:
- RSIR offers an efficient and stable approach for TFBM identification.
- The method improves the sensitivity and specificity of motif detection by avoiding model misspecification.