Fig. 4: Estimating ligation efficiencies.

a Products of single and dual ligations of the RBP hnRNP C; fluorescent markers ligated to RNA (left three panels) or a western blot of cell lysate (right panel). “Larger complexes” contain ligated adapter and hnRNP C protein; they likely reflect the presence of additional proteins, but their exact nature is unknown. The experiment was performed three times with similar results. IB: immunoblot. b Chromatographs generated from the fluorescence gel image in panel (a), far-right lane. c Calculations to determine total bound RNA from the three observed values and the assumption of statistical independence in ligation efficiencies to determine the fourth value. d Ligation efficiencies estimated by the protein shift method (n = 7), by RNA shift (n= 4), or by shifting free adapters (n= 3). Center line, median; bounds of box, 25–75% quartile range; whiskers, maxima, and minima. e Ligation efficiency by shifted RNA. The experiment was performed four times with similar results. f Correlation between ligation efficiencies determined by the protein shift method or the RNA shift method for four biological replicates. g Five representative RBPs have 5′ ligation efficiencies of 42–95%, as determined by protein shift. DDX50 binds rRNA, while the others are mRNA-binding proteins. hnRNP D binds AU-rich RNA, CELF binds UGU, hnRNP C binds poly(U), and SRSF2 binds a GA-rich motif. Fifty percent is a reasonable approximation for the 5′ ligation efficiency in general. Data are mean ± 95% CI. h The 5′ ligation efficiency is robust to RNAse concentration (n = 3 independent samples). RNAse concentrations of 0.05–5.0U/µL resulted in <2-fold changes in ligated RNA. This is likely due to the RNAse and the ligase having similar limits on RNA length for a 5′ fragment. Bars represent the mean. AU arbitrary units. i Diagram of the method of absolute RNA quantification. The amount of observed ligated RNA is multiplied by two based on the observed general ligation efficiency in order to determine the number of cross-linked RNA molecules. When a quantitative immunoblot is performed on the same gel for the uncross-linked protein, the RNA molecule count may be divided by the protein molecule count to derive the cross-link rate. j Method to combine sequencing data and cross-link rates to determine the cross-links to an RNA (or class of RNAs, region of RNAs, etc.) per protein molecule.