Blocking and Randomization to Improve Molecular Biomarker Discovery

• Activity Number: 603
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Li-Xuan Qin*+ and Qin Zhou and Jaya Satagopan and Colin Begg and Douglas Levine
• Companies: Memorial Sloan Kettering Cancer Center and Memorial Sloan Kettering Cancer Center and Memorial Sloan Kettering Cancer Center and Memorial Sloan Kettering Cancer Center and Memorial Sloan Kettering Cancer Center
• Keywords: microrna ; microarray ; normalization ; randomization ; blocking ; experiment design

Abstract

Background: Randomization and blocking can prevent the negative impact of non-biological effects in genomic studies of biomarkers. Their use in practice, however, has been scarce.

Methods: To demonstrate the logistic feasibility and scientific benefits of randomization and blocking, we conducted a microRNA study of endometrial tumors (n=96) and ovarian tumors (n=96) using the blocked randomization design. The same set of tumors was profiled for a second time using no blocking or randomization. We assessed empirical evidence of differential expression in both studies. We conducted simulation studies to further evaluate the effects of randomization and blocking.

Results: There was moderate and asymmetric differential expression (10%=351/3523) between endometrial and ovarian tumors in the randomized dataset. Array effects were observed in the non-randomized dataset and 1934 markers (55%) were called to be differentially expressed. Among these markers, 181 were deemed DE (181/351, 53%) and 1749 non-DE (1749/1934, 90%) in the randomized dataset. In the simulation study, blocking with randomization had the best accuracy of biomarker detection (TPR=0.97, FPR=0.001), followed by random