Supplementary Figure 4: Using HT-KAM as a discovery platform to identify best targets of kinases, establishes that biological peptides are systematically found as reporters of the most significantly high activity profiles of their respective kinases.

(a) Binary heatmap of peptides associated with most significantly high/low activity per kinase. Computational processes and statistical cut-offs comparing levels of ATP consumption per individual peptide to the pool of reference peptides for each kinase are described in Methods. This stringent selection finds 110 biological peptides act as robust sensors of kinases’ catalytic activities. (b) Analysis of activity profiles established in presence of inhibitors (n=3 per treatment condition). Here, the underlying postulate is that, when the activity of a kinase is measured in presence of an inhibitor, any peptide associated with a significant decrease in activity may be considered as a suitable sensor to detect the activity of this kinase. Results would assess the utility and reliability of kinases’ biological peptide targets and peptides defined in (a). For instance, for ABL1, quadrant A shows a strong correlation between levels of inhibition (y-axis; Pearson correlation coefficient between imatinib concentration and ATP consumption for each peptide), and the activity level per peptide in an untreated context (x-axis) (R²(Fisher(inhibition), activity)=-0.48; p=2.75e-14). Critically, peptides that report higher ABL1 activity levels (i.e. dots located toward the right-end of the x-axis, indicating highest ATP consumption) exhibit greater activity inhibition in presence of increasing concentrations of imatinib (i.e. dots located toward the bottom-end of the y-axis, indicating strongest negative correlation and thus strongest inhibition). These peptides largely overlap with peptides found in (a) (i.e. red dots located in the bottom-right area in quadrant A), and include ABL1’s biological peptides JUN Y170, CDK5 Y15, WASL Y256 or BTK Y223, confirming that ABL1 phosphorylation of its substrates is targetable by anti-ABL1 therapy. This analysis also finds that the other biological peptides of ABL1 (i.e. TP73 Y99, JAK2 Y1007, MAP4K1 Y232, ABL1 Y226, CDKN1B Y88, MDM2 Y394, RAD51 Y54; see Fig. 3a) are systematically and significantly associated with ABL1 kinase activity and measurable response to ABL1-inhibitors. This is further validated with results from dasatinib-treated ABL1 (quadrant B). The same logic applies to LYNA (quadrant C), and can be applied to any kinase, peptide set, and drug. (c-d) Comparison of most repeatable peptide-derived activities of ABL1 (x-axis; defined in (a)) to the spectrum of activities established from ABL1’s differential signature (y-axis; defined in Supplementary Fig. 2g–i). Areas (i)/(ii) in (c) and (d) match. Red-filled/outlined markers correspond to biological peptides of ABL1. Noticeably, all ABL1 biological peptides report on high ABL1 activity (top right corner in (c)). Peptides associated with significantly low ABL1 activity correspond to biological peptides of Ser/Thr-kinases (listed in (d)). This directly supports why libraries combining ‘disparate’ biological peptides are inherently good discriminators and identifiers of different kinases. (e-f) Venn diagrams intersecting results from analysis in (a) (robustness of peptide-phosphorylation activities) and in Fig. 1e, f and Supplementary Fig. 2h-i (most differential peptide activities). Results similar to ABL1 in (e) are found for the other kinases we tested (e.g. AKT1 in (f)). In all cases, biological peptides act as robust sensors of the differential and individual/specific phospho-catalytic activity signatures of kinases. (g) Comparing phospho-catalytic profiles of AKT1 versus AKT2. Activity differences are more significantly measurable with biological peptides than generic CON⁺ peptides, also revealing exploitable differences (e.g. AKT2-signaling networks). Based on results obtained from computational methods we developed and combinatorial peptide library system we designed to showcase our strategy, one valuable application of the HT-KAM approach is to serve as a discovery platform to find new biological substrates and signaling relationships to generate testable hypotheses and explore uncharted druggable pathways.