The PCA analysis of YRO time series has shown that a reduction of the data to standardized cohort means still retrieves experiment sampling times as the phase angle spanned by the first two eigenvectors atan2(PC2,PC1). Here, we use both the YRO cohorts and cell division cycle-related cohorts (Spellman et al. 1998) as temporally ordered sets of marker genes to sort single cells by their expression status.
Expectedly, the picture in single cell data (Gasch et al. (2000), Jackson et al. (2020), (unpublished data by Ben Heineike) is more complex and 2 dimensions seem to fail to clearly align YRO and CDC expression. However, a global trend becomes visible such that the cohort is associated with M and G1 phases, while the cohort is co-initiated in S phase but remains active throughout G2. The amplitudes of this pattern strongly increase at slower growth rates (minimal and limited media), but are present in all growth conditions, including full medium (YPD) and the non-growing condition (CStarve).
Here are movies of the first three dimensions, for PCA of standardized means of the