Figure 6.

Genes that show more than twofold induction in the HAP4 overexpression dataset can be divided in two distinct groups using k-means clustering analysis (Cluster software [22]). The gene-expression profile of HAP4-overexpressing cellswas compared to datasets derived from cells that have a particular and/or shifted balance between respiration and fermentation. These were: the 18.5 h and 20.5 h time points of the diauxic shift at which glucose has been exhausted [1]; the 'adaptive evolution' datasets, which are derived from strains selected for growth in aerobic, glucose-limited chemostats for more than 250 generations (denoted as pare1, pare2 and pare3 [28]); stationary phase cells (2-day stationary phase); and expression profiles from cells grown either on ethanol or glucose [11]. It should be emphasized that mitochondrial transcripts are not represented in this analysis as these transcripts do not contain a 3' poly(A) tail and hence are not labeled properly.