93 (Phospho)proteomics biomarkers for sunitinib response in RCC a more complete data matrix, a stricter filter could be applied. No imputation of data was performed. Heatmap visualization and hierarchical clustering was done with the R package ComplexHeatmap version 2.2.039. Differential proteins were imported into Cytoscape version 3.540, and gene ontology analysis was performed in Cytoscape with the BiNGO app version 3.0.341, using ontology and organism annotation definitions downloaded on 8 July 2019 via http://geneontology.org. KINASE ACTIVITY ANALYSIS Per sample, a ranking of most activated kinases was generated using the Integrative Inferred Kinase Activity (INKA) data analysis pipeline24, taking both information on phosphorylated kinases and their substrates into account. Differentially activated kinases were identified and level of significance was determined by Mann-Whitney U-test. POST-TRANSLATIONAL MODIFICATIONS SIGNATURE ENRICHMENT ANALYSIS (PTM-SEA) PTM-SEA42 was performed using the Phospho (STY).txt Max Quant search result file after filtering out decoy and contaminant site entries, to identify site-specific signatures of kinase activities and signaling pathways, overrepresented in each of the 2 groups. Phosphosites were ranked using -10 * sign(logFC) * log10(P-Value) as a measure, where the P-value and logFC were calculated in a differential analysis by limma version 3.38.3. and used as inputs to run the PTM-SEA algorithm in GenePattern43 (https://cloud.genepattern.org). The PTM signature sets were those defined in PTMsigDB v1.9.0 (human, flanking sequence format, file ptm.sig.db.all. flanking.human.v1.9.0.gmt) downloaded from https://github.com/broadinstitute/ssGSEA2.0. Results were visualized in R. Significantly enriched signatures were reported (FDR < 0.25). EXPLORATION OF (PHOSPHO)PROTEOMICS CANDIDATES IN TRANSCRIPTOME DATA OF AN INDEPENDENT COHORT Publicly available transcriptomics data from an independent cohort previously described by Beuselinck et al44 was used. CEL files containing Affymetrix array signals from 59 patients with ccRCC, treated with sunitinib, were obtained and processed in R (package “oligo”). Group comparison analysis was done in R (package “LIMMA”). All significantly (p < 0.05) differentially expressed transcripts were considered. Expression levels of differentially expressed proteins from our proteomics analysis (p < 0.05 & FC > 2 & ≥ 50% data points in the highest group) were compared to the expression of matching transcripts in the validation cohort at gene level, the percentage of overlapping proteins/transcripts was reported. DATA AND MATERIALS AVAILABILITY The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE45 partner repository with the dataset identifier PXD043514. 4
RkJQdWJsaXNoZXIy MTk4NDMw