Figure 1: Integrative analysis of omics data reveals targetable kinases in NSCLC KRAS-Dep cell lines.

(a) A panel of KRAS-Dep and KRAS-Ind cell lines was analysed by transcriptomics, proteomics and phospho-proteomics techniques. Transcripts were split in two different categories: ‘informative’ genes and ‘all other’ genes. Proteome and phopho-proteome data sets were normalized with respect to the total number spectral counts in each library, and common contaminants and ‘Deja vu’ proteins were filtered out before quantification of differential abundance. All data sets were log-transformed and the LFC was taken with respect to the comparison KRAS-Dep versus KRAS-Ind cell lines. The LFC was z-score-normalized and a P-value was calculated using the standard normal distribution. The combined S score was used to integrate all three data sets (methods) and select differentially expressed proteins. Network and enrichment analysis were performed using SPIA and the PCST. (b)Naive integration of data sets. Only~5.2% of the proteins are shared among two of the data sets (adjusted P-value 0.05 was used as a threshold to select differentially expressed proteins). A major drawback of this method is the absence of an objective criterion to include proteins differentially expressed in only one data set. (c)A meta-integration of the independent signatures using the combined S score (S). The S score integration improves by fivefold the percentage of shared proteins among data sets (~26%) and defines an objective rule for including proteins differentially expressed in one, two or all data sets. (d)Integrative analysis of transcriptome, proteome and phospho-proteome nominates receptor tyrosine kinases MET and ERBB3, Src family members LCK and LYN, PAK1, and CTNNB1, CTNNA1 and CDH1 among others as differentially ‘activated’ proteins in KRAS-Dep cell lines. Left: presence/absence heatmap. Proteins that are differentially abundant in a particular data set are represented in yellow and unaffected proteins are represented in blue. Middle: combined S score (S) for all differentially abundant proteins in KRAS-Dep versus KRAS-Ind cell lines. Right: combined statistical significance for each differentially abundant protein. −log of the Hochberg-adjusted P-value, −log(0.05)=1.30.