Extended Data Fig. 6: TME binders cover disordered proteins of low abundance.

a-c, Violin (left two panels) and bar (right two panels) plots for “Lysate,” “Binder,” and “Undetectable” in the MG132 dataset, using Neuro-2a cell line proteomic dataset (a), the PaxDb Brain (Integrated) dataset (b), and the PaxDb Whole Organism (Integrated) dataset (c) as standard references. Violin plots show the distribution of protein abundance (n = 2893 for Lysate, 680 for Control Binder, 107 for Control Undetectable, 697 for MG132 Binder, 88 for MG132 Undetectable, to be mapped to the corresponding reference dataset; P value was calculated using the unpaired two-sample two-sided Wilcoxon test; ****, P < 0.0001). The boxplot displays five key elements: the median, shown as the white circle at the center; the upper hinge, representing the third quartile (75th percentile); and the lower hinge, representing the first quartile (25th percentile). The upper whisker extends to the largest value within 1.5 * IQR (interquartile range) from the upper hinge, while the lower whisker extends to the smallest value within 1.5 * IQR from the lower hinge. Bar plots illustrate the percentages of proteins within each percentile, with percentiles determined by the respective standard references. ‘U’ in the last column of the bar plot denotes ‘Unmapped,’ referring to proteins not found in the corresponding standard references. “Lysate,” “Binder,” and “Undetectable” denote: proteins detected by lysate profiling methods, TME binders identified by RUBICON exclusive of undetectable proteins, and TME binders only detected by RUBICON but not by lysate profiling methods (red dots in Extended Data Fig. 5a-b), respectively. Notably, referencing the Cell Line dataset results in the most unmapped proteins, followed by the Brain dataset, which leaves some proteins unmapped. Using the Whole Organism dataset, nearly all protein abundances can then be mapped.