Contents     Editorial   Commentary Research Glossary Feedback

Using RNA sample titrations to assess microarray platform performance and normalization techniques

Richard Shippy, Stephanie Fulmer-Smentek, Roderick V Jensen, Wendell D Jones, Paul K Wolber, Charles D Johnson, P Scott Pine, Cecilie Boysen, Xu Guo, Eugene Chudin, Yongming Andrew Sun, James C Willey, Jean Thierry-Mieg, Danielle Thierry-Mieg, Robert A Setterquist, Mike Wilson, Anne Bergstrom Lucas, Natalia Novoradovskaya, Adam Papallo, Yaron Turpaz, Shawn C Baker, Janet A Warrington, Leming Shi & Damir Herman

Supplementary Fig. 1 (doc 775K)
Linear regression fitting of C_i/B_i vs. A_i/B_i for each platform, normalization method and site to estimate the true mixture coefficients _c,_c for the C sample from the data.

Supplementary Fig. 2 (doc 60K)
Illustration providing a possible explanation for the asymmetry (i.e., median _A>B (C) median_A<B (C) and median _A>B (D) median_A<B (D)) in the box plots on either side of Figure 4 (particularly the differences in median values).

Supplementary Fig. 3 (doc 60K)
Percentage of genes achieving the specified level of apparent power between the RNA titration mixtures and independent samples.

Supplementary Fig. 4 (doc 430K)
Impact of different normalization methods on signal CVs for each microarray platform separated by site and sample.

Supplementary Fig. 5 (doc 45K)
Derivation of formulas within Box 1 to clarify calculations for mRNA fractions.

Supplementary Table 1 (doc 45K)
Gene counts for each commercial whole genome microarray platform separated by site.

Top