Fig. 3
From: High-throughput discovery of post-transcriptional cis-regulatory elements
Assaying 8mers with known regulatory properties. a Regulatory effects of expression controls. Each control sequence was inserted into the IQGAP1 3′UTR within a luciferase reporter. The luciferase activity of these reporter constructs were normalized to the geometric means of two control sequences with no regulatory effect (UUCCGUUA and UAAUGCCC). The normalized geometric means are plotted ± 33 % of the spread of the data. Significance was determined by two-sided Wilcoxon rank sum tests; n = 9. * p < 0.05, ** p < 0.005, *** p < 0.0005. b Expression of control sequences when inserted into the integrated GFP reporter. The distribution of GFP intensity is shown for each control as compared to unsorted cells. c Correlation of relative luminescence and GFP values. The geometric mean of the GFP intensities for each expression control were calculated and normalized to the geometric means of two control sequences with no regulatory effect (UUCCGUUA and UAAUGCCC). These values (y-axis) were plotted against the relative luminescence values (x-axis) found in B. Pearson r = 0.975, p < 10−5. d Performance of expression controls in the main screen. Cells containing each of the ten expression controls were added to the library of cells prior to sorting. DNA was isolated from the sorted populations, and the 8mers were identified and quantified. Read values were normalized by the number of reads in each sequencing library (reads per million, RPM). The enrichment of each expression control was calculated by dividing the RPM values in each GFP sub-population by the RPM value in GFP-ALL cells. e Correlating enrichment in sorting bins with an 8mer’s GFP intensity. Each control’s relative GFP intensity (x-axis) is plotted against enrichment in the five sorting bins (y-axis), and the lines of best fit are shown. Correlation values (Pearson r) for each bin against intensity are as follows: 0–10 % (r = −0.978, p < 10−5); 20–30 % (r = −0.863, p < 0.005); 40–60 % (r = 0.369, not significant); 70–80 % (r = 0.787, p < 0.05); 90–100 % (r = 0.972, p < 10−5)