Viral fitness predicts the magnitude and direction of perturbations in the infected host transcriptome

Héctor Cervera; Silvia Ambrós; Guillermo P. Bernet; Guillermo Rodrigo; Santiago F. Elena

erence genes and then averaged. Finally, to make expression data by both methods (microarray readings and RT-qPCR) readily comparable, they were both transformed into z- For each gene, the right plot shows both expression z-scores as a function of TEV fitness; solid lines represent the best linear fitting between normalized expressions and TEV fitness. In this representation, the more overlap between the two regression lines, the better the agreement between both quantitative methods. In this representation, VQ29 and ADK2 show the largest departure between both regression lines, thought even in these extreme cases, the difference was not large enough as to be significant in a non-parametric Wilcoxon's signed ranks test (P ³ 0.499 in all nine cases) or in a Student's t-test for the comparison of regression coefficients (P ³ 0.285).