What statistical parameter can be applied to a gene expression study after REST analysis?

Real time PCR, Gene Expression


0 / 0  ·  3 Answers  ·  158 Views

All Answers (3)

  • Veeresh Sangappa · Institute of Bioinformatics and Applied Biotechnology
    Real time PCR
  • Isacco Beritognolo · National Research Council
    the results of REST statistical analyisis include the significance level ("p(H1)") of the relative expression ratio ("expression"). The parameter p indicates the probability of 1st type error when accepting the hypothesis that the relative expresssion ratio is different from zero (i.e. UP- or DOWN-regulated). As default rest considers the result statistically significant when p<0.05, but you could use a more stringent threshold, for example p<0.01.
  • Maciej Woźny · University of Warmia and Mazury in Olsztyn
    Good question. REST software gives you the ability to perform the Randomisation Test, which in my opinion is very convenient statistical approach in gene expression analysis. However, the construction of this computer program allows only for simple comparisons of two experimental groups (e.g. control vs. treated) and it's not enough for multiple comparisons. In this case, you could try to perform additional parametric tests, but then keep in mind that the gene expression data should fulfill the appropriate requirements (e.g. distribution, homogeneity of variance) which is hard to obtain without any transformations.

    Hoping to read any other propositions. Cheers!

Question Followers (6)