(A) Distributions of signal from the GFP reporter (in arbitrary units), measured from FACS, in cells expressing artificial TF (aTF) fused to ADs from VP16, Gcn4, or Pho4, or without an AD. Cells were filtered by a narrow range of mCherry, which tracks the aTF expression level. (B) FACS plots of 1,000,000 events observed in the activation assay with the TF tiling and AD mutant sub-libraries (sub-libraries A and B), showing (left) the mCherry and GFP signal for each cell, (middle) the mCherry range used for sorting, and (right) the distribution across the eight GFP bins. (C) In vivo fold-activation and corresponding Z-scores measured for 10 positive control ADs that were each encoded by eight different synonymous DNA sequences. Horizontal axis labels show the protein and start position of each 53-aa AD. (D) Distributions across the eight GFP bins of cells expressing aTF fused to ADs from VP16, Gcn4, or Pho4, either measured individually by FACS (red line) or in the pooled screen encoded by eight different synonymous DNA sequences (black lines). (E) mCherry versus GFP and GFP versus GFP/mCherry ratio for cells expressing aTF with no AD or with the VP16 AD. The GFP/mCherry ratio alone poorly distinguishes GFP-positive cells. (F) aTF expression is affected by tile sequence. To estimate how tile protein sequences affected aTF expression levels, we sequenced theFACS input cells and examined, for each sequence, the frequency at which input cells fell within the filtered mCherry range (i.e. were seen across all eight GFP bins combined). Plotted here is the mean (black) and standard deviation (gray bars) of this measure for the hydrophobic mutants of the Gcn4 AD (Figure 2D), relative to wild-type, when binned by their hydrophobic content. (G) Scatter plot of activation of tiles from 150 ADs (green) and 50 random sequence controls (blue) shared between two separately cloned and assayed libraries (sub-libraries A and D, Pearson's r = 0.952). (H) ADs were annotated using the (log-scale) mean activation at each protein position, shown here for Adr1. Width of ADs were determined by the full width half maximum of each peak that crossed the threshold (dashed line). Annotated ADs are marked in green bars. Related to Figure 1D–E. (I) (Top) C-terminal ADs are in nearly half of all AD-containing TFs. The horizontal axis shows protein position normalized so that the N- and C-termini are at 0 and 1, respectively. (Bottom) Cumulative distribution plots of the length and maximal activation (among overlapping tiles) for ADs at the C-terminus of a protein (green) versus other ADs, showing that C-terminal ADs are shorter and stronger on average. (J) TFs that contained ADs upregulated a higher proportion of downstream genes than TFs without ADs, p=1.5E-7 by KS test. Genes up- or down-regulated by each TF were annotated by Hackett et al., 2020.