Table 3 Involvement of tumor epithelium and stroma. Linear regression was used to quantify the relative contributions of tumor epithelium and stroma to the gene expression profile of the macrodissected sample. If, for one gene, s is the amount of RNA measured in microdissected stroma, t is the amount measured in microdissected tumor, and r is the amount RNA measured in the macrodissected sample, we assume r = αt+βs, where α and β are unknown coefficients. The last column is the relative contribution of tumor epithelium, α/(α+β), assuming that the contributions of stroma and tumor together are 100%. Because we are considering a sum of contributions on the linear RNA scale, the regression has to be performed on the non-logged data. The values are averaged in case of duplicate labeling experiments and standard errors of coefficients α and β are given.

From: Macrodissection versus microdissection of rectal carcinoma: minor influence of stroma cells to tumor cell gene expression profiles