Fig 1.
A minimalistic model predicts that low expression of condition-dependent genes under a permissive growth environment, compared with a restrictive environment, implies larger fraction of all other proteins out of the proteome.
With this, the ratio of bio-synthesis genes to the rest of the proteome is higher in permissive environments, resulting in faster growth.
Fig 2.
A strong positive Pearson correlation between the fraction out of the proteome and the growth rate is observed for a large number of proteins in two data sets.
(A-B) Shown are histograms displaying the correlations of all proteins to growth rate in the data from [29] (A) and [13] (B). Functional protein groups are denoted by different colors. Thresholds defining high correlation are marked in dashed lines and further discussed in S4 Text. (C) Shuffling the amounts of every protein across conditions for the data set from [13] reveals the bias towards positive correlation with growth rate is non-trivial.
Fig 3.
Histogram of the slopes of regression lines for the highly correlated with growth proteins (473 and 305 proteins in the left and right panels respectively).
Ribosomal proteins are stacked in green on top of the non ribosomal proteins, marked in blue. Proteins fractions were normalized to account for differences in slopes resulting from differing average fractions (Methods). The expected distribution of slopes given the individual deviations of every protein from a linear regression line, assuming all proteins are coordinated, is plotted in gray. Dashed vertical lines at 0.5 and 2 represent the range at which the slopes of of the proteins lie. Left panel—data from [29], right panel—data from [13]. High correlation proteins share similar normalized slopes, implying they are coordinated, maintaining their relative ratios across conditions (see text for further details). Ribosomal proteins, shown in green, scale with growth rate in a manner similar to the rest of the high correlation proteins (see text and S7 Fig).
Fig 4.
Fraction of the proteome occupied by proteins that are strongly positively correlated with growth rate.
The accumulated sum of the proteins that are strongly positively correlated with growth rate (defined as having a correlation above 0.5), as a fraction out of the proteome, with linear regression lines is shown. These proteins form a large fraction (≥ 50%) out of the proteome at higher growth rates. The accumulated fraction of the strongly correlated proteins doubles as the growth rate changes by about 5-fold. Assuming constant degradation rates, the trend lines correspond to protein half life times of ≈ 1.7 hours. Randomized data sets result in much fewer strongly positively correlated with growth rate proteins, implying a much smaller accumulated fraction (hollow circles).