Supplementary Figure 7: Characterizing the phospho-catalytic signatures of melanoma tumors is relevant for patient diagnostic and treatment guidance.

(a) Schematic of the procedure. (b) Baseline ATP levels measured between experiments and tumors. (c) Unsupervised hierarchical clustering of peptide phosphorylation signatures of nine patient tumor biospecimens, tested in four independent replicates, and measured across 228 peptides. (d) Peptide phosphorylation profiles using top-25%-most-differential activities (alternative threshold-selection method complementing Fig. 6c, d). The related differential activities per peptide comparing specific patient groups are available in Supplementary Table 21. Results in (d) and Fig. 6c, d show that most significant and consistently high signals associated with fatal outcome were measured with biological peptides SMAD2 S465/S245/S250, KHDRBS1 Y440/Y435, MTOR T2446/S2448, CDKN1A T145/S146, BRCA1 S988/T509, ABL1 Y226 (i.e. Y245 in another ABL1 isoform), FCGR2B Y292, CHEK1 S280. Peptides displaying consistently low activity in fatal outcome patient tumors were biological peptides NOTCH2 S2070, JUN Y170, TERT Y707, GAB1 Y627, or reference/modified biological peptides PA_128, PA_134, PA_230. Poor outcome-associated high-activity signatures included peptides related to kinases such as MTOR, SFKs, or TGFBR-signaling, some of which conferred acquired resistance to BRAF-/MEK-inhibitors in melanoma cell lines ^45,64-66. Activities measured with peptides of related amino-acid sequences behaved similarly, serving as internal controls validating repeatability (e.g. MTOR T2446/S2448; CDKN1A T145/S146; CDK5 Y15 and CON⁺ PA_240 derived from CDK1 Y15 conserved phospho-site). Thus, peptides can be used as robust phosphorylation activity sensors to generate clinically valuable signatures. (e) Kinase signature of tumors. (f) Waterfall plot showing kinases’ differential activities between patient groups. Some kinase activities identified by HT-KAM in (e-f) are corroborated by a gene over-expression study of RAF-inhibitor resistance in a melanoma cell line model ⁴⁵, including upregulation of PRKCE, cRAF, MAP3K8/COT1 or PAK1, although the activity of these kinases were not as highly and significantly upregulated as PIM, RPS6KB or AKT in poor outcome/VEM-resistant patient tumors. (g) Comparing ABL1, AKT1, ERK2, p38a, HCK activities deconvoluted from kinases’ biological peptide subsets (x-axis) versus derived from kinases’ differential peptide sets (y-axis; peptides identified in Supplementary Fig. 2g–i) across different patient groups. Good concordance further validates our strategy. (h) Differential biological peptide-phosphorylation activity profiles complementing Fig. 7b. Results in (e-h) and Fig. 7a, b demonstrate that HT-KAM can measure differences in kinases’ activities (i) within a tumor biospecimen (i.e. for each patient), (ii) between different tumors (i.e. different patients), (iii) between different patient groups (e.g. outcome), which can reveal significantly hyperactive kinases predictive of survival, and identify druggable kinase vulnerabilities most tractable to treat patients unresponsive to BRAF-therapy and with highest likelihood of recurrence.