|
Activity Number:
|
609
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #303880 |
|
Title:
|
Inferring the True Correlation in Cross-Species Microarray Data
|
|
Author(s):
|
George C. Tseng and Xingbin Wang*+ and Sunghee Oh
|
|
Companies:
|
University of Pittsburgh and University of Pittsburgh
|
|
Address:
|
, , ,
|
|
Keywords:
|
microarray ; correlation ; comparative genomics
|
|
Abstract:
|
Correlation measures are often used to quantify gene effect correlations across microarray studies. Poor correlations are often found in related studies of the same disease or cross-species comparison and biological and experimental variations are the major causes. Here we hypothesize that poor correlations are also partially caused by the mixing of many non-correlated and inactive genes. We propose to apply a extreme-value correlation (EVC) measure where the correlation is calculated on genes of large absolute effects. A weighted least square approach is used to correct the bias from EVC and to estimate the underlying correlation via a contaminated Gaussian mixture model. The hypothesis and methods are validated by simulation and by a microarray data on human aging versus mouse aging comparison. The result shows improved bias and efficiency over naive Pearson correlation.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2009 program |
