Related Experiment Videos
Statistical methods of translating microarray data into clinically relevant diagnostic information in colorectal
Byung Soo Kim1, Inyoung Kim, Sunho Lee
1Department of Applied Statistics, College of Medicine, Yonsei University Seoul, South Korea. bskim@yonsei.ac.kr
Bioinformatics (Oxford, England)
|September 18, 2004
Summary
This study introduces a new statistical method to analyze mixed microarray data from normal and tumor tissues. The developed Hotelling's T2 statistic efficiently detects differentially expressed genes, improving upon standard procedures.
Area of Science:
- Bioinformatics
- Genomics
- Cancer Research
Background:
- Cancer microarray experiments often collect paired normal and tumor tissues.
- Experimental limitations can result in mixed datasets with paired and independent samples.
- Standard statistical methods are insufficient for analyzing such mixed microarray data.
Purpose of the Study:
- To develop a novel statistical method for analyzing mixed microarray datasets from cancer studies.
- To identify differentially expressed genes between normal and tumor tissues using combined data.
Main Methods:
- Proposed a new test statistic, t3, to integrate information from mixed datasets.
- Utilized the extended receiver operating characteristic approach.
- Employed Hotelling's T2 statistic for detecting differentially expressed genes.
Main Results:
- The proposed t3 statistic effectively combines information from mixed sample types.
- Hotelling's T2 statistic demonstrated higher efficiency in detecting differentially expressed genes compared to univariate methods.
- A measure of disagreement between RT-PCR and microarray experiments was devised.
Conclusions:
- The developed statistical approach is suitable for analyzing mixed microarray data in cancer research.
- Hotelling's T2 statistic offers a more efficient method for identifying differentially expressed genes.
- Further research into formal test procedures using Hotelling's T2 statistic is recommended.