Reverse phase protein arrays (RPPA) are an efficient, high-throughput, cost-effective method for the quantification of specific proteins in complex biological samples. The quality of RPPA data may be affected by various sources of error. One of these, spatial variation, is caused by uneven exposure of different parts of an RPPA slide to the reagents used in protein detection. We present a method for the determination and correction of systematic spatial variation in RPPA slides using positive control spots printed on each slide. The method uses a simple bi-linear interpolation technique to obtain a surface representing the spatial variation occurring across the dimensions of a slide. This surface is used to calculate correction factors that can normalize the relative protein concentrations of the samples on each slide. The adoption of the method results in increased agreement between technical and biological replicates of various tumor and cell-line derived samples. Further, in data from a study of the melanoma cell-line SKMEL-133, several slides that had previously been rejected because they had a coefficient of variation (CV) greater than 15%, are rescued by reduction of CV below this threshold in each case. The method is implemented in the R statistical programing language. It is compatible with MicroVigene and SuperCurve, packages commonly used in RPPA data analysis. The method is made available, along with suggestions for implementation, at http://bitbucket.org/rppa_preprocess/rppa_preprocess/src.

RPPA analysis from FAIRLANE Trial of Neoadjuvant Ipatasertib Plus Paclitaxel for Triple-Negative Breast Cancer.

Pancreatic cancer cells MIAPaca2 under mechanical compression

The columns represent protein levels of phospho to total ERK ½ (A), Rho GDI (B), caldesmon (C), nucleolin (D) and SUMO1 (E) in control (blue), 1.4 mGy/h (orange) and 2.4 mGy/h (green) irradiated HUVECs. The average ratios of relative protein expression in control and irradiated samples are shown. The data are represented as ± SEM. Three biological replicates were used in all experiments. (Students t-test; *p

The SUM human breast cancer cell lines have been used by many labs around the world to develop extensive data sets derived from comparative genomic hybridization analysis, gene expression profiling, whole exome sequencing, and reverse phase protein array analysis. In a previous study, the authors of this paper performed genome-scale shRNA essentiality screens on the entire SUM line panel, as well as on MCF10A cells, MCF-7 cells, and MCF-7LTED cells. In this study, the authors have developed the SUM Breast Cancer Cell Line Knowledge Base, to make all of these omics data sets available to users of the SUM lines, and to allow users to mine the data and analyse them with respect to biological pathways enriched by the data in each cell line.Data access: All the datasets supporting the findings of this study are publicly available in the SLKBase platform here: https://sumlineknowledgebase.com/. RPPA data, drug sensitivity data, apelisib response data, and data on dose response, are also part of this figshare data record (https://doi.org/10.6084/m9.figshare.12497630).Study aims and methodology: This web-based knowledge base provides users with data and information on the derivation of each of the cell lines, provides narrative summaries of the genomics and cell biology of each breast cancer cell line, and provides protocols for the proper maintenance of the cells. The database includes a series of data mining tools that allow rapid identification of the functional oncogene signatures for each line, the enrichment of any KEGG pathway with screen hit and gene expression data for each of the lines, and a rapid analysis of protein and phospho-protein expression for the cell lines. A gene search tool that returns all of the functional genome and functional druggable data for any gene for the entire cell line panel, is included. Additionally, the authors have expanded the database to include functional genomic data for an additional 29 commonly used breast cancer cell lines. The three overarching goals in the original development of the SLKBase are: 1) to provide a rich source of information for anyone working with any of the SUM breast cancer cell lines, 2) to give researchers ready access to the large genomic data sets that have been developed with these cells, and 3) to allow researchers to perform orthogonal analyses of the various genomics data sets that we and others have obtained from the SUM lines. For more information on the development and contents of the database, please read the related article.Datasets supporting the paper:The data mining tools accessed the following datasets to generate the figures and tables, and these datasets are downloadable from the Data Download centre on the SLKBase: Exome sequencing data: SLKBase.exome_.seq_.sum_.xlsxGene amplification and expression data for the SUM cell lines: SUM44amplificationdata.xlsSUM52.xlsSUM149.xlsSUM159.xlsSUM185.xlsSUM190.xlsSUM225.xlsSUM229.xlsSUM1315.xlsCellecta shRNA screen data for the SUM cell lines:SUM44Celectadata.csvSUM52Cellectadata.csvSUM102Cellectadata.csvSUM149Cellectadata.csvSUM159Cellectadata.csvSUM185Cellectadata.csvSUM190Cellectadata.csvSUM225Cellectadata.csvSUM229Cellectadata.csvSUM1315hits.hit.csvMCF10A.hits_.csvBreast cancer cell line data included in this data record (these datasets were used to generate figures 1, 2 and 7 in the article):Proteomics data from the Reverse Phase Protein Array (RPPA) assay analysis: Ethier.SUMline.RPPA.xlsxDrug sensitivity data: NAVITOCLAX.drugsensitivity.Zscores.xlsxApelisib response data: Apelisib all lines (2).xlsxDose response data: 092614 Dose Response CP 52s.11.15.xlsxAll the files are either in .xlsx or .csv file format.

Integrating data from multiple regulatory layers across cancer types could elucidate additional mechanisms of oncogenesis. Using antibody-based protein profiling of 736 cancer cell lines, along with matching transcriptomic data, we show that pan-cancer bimodality in the amounts of mRNA, protein, and protein phosphorylation reveals mechanisms related to the epithelial-mesenchymal transition (EMT). Based on the bimodal expression of E-cadherin, we define an EMT signature consisting of 239 genes, many of which were not previously associated with EMT. By querying gene expression signatures collected from cancer cell lines after small-molecule perturbations, we identify enrichment for histone deacetylase (HDAC) inhibitors as inducers of EMT, and kinase inhibitors as mesenchymal-to-epithelial transition (MET) promoters. Causal modeling of protein-based signaling identifies putative drivers of EMT. In conclusion, integrative analysis of pan-cancer proteomic and transcriptomic data reveals key regulatory mechanisms of oncogenic transformation.

Pancreatic cancer is one of the most lethal tumors, and reliable detection of early-stage pancreatic cancer and risk diseases for pancreatic cancer is essential to improve the prognosis. As 260 genes were previously reported to be upregulated in invasive ductal adenocarcinoma of pancreas (IDACP) cells, quantification of the corresponding proteins in plasma might be useful for IDACP diagnosis. Therefore, the purpose of the present study was to identify plasma biomarkers for early detection of IDACP by using two proteomics strategies: antibody-based proteomics and liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based proteomics. Among the 260 genes, we focused on 130 encoded proteins with known function for which antibodies were available. Twenty-three proteins showed values of the area under the curve (AUC) of more than 0.8 in receiver operating characteristic (ROC) analysis of reverse-phase protein array (RPPA) data of IDACP patients compared with healthy controls, and these proteins were selected as biomarker candidates. We then used our high-throughput selected reaction monitoring or multiple reaction monitoring (SRM/MRM) methodology, together with an automated sample preparation system, micro LC and auto analysis system, to quantify these candidate proteins in plasma from healthy controls and IDACP patients on a large scale. The results revealed that insulin-like growth factor-binding protein (IGFBP)2 and IGFBP3 have the ability to discriminate IDACP patients at an early stage from healthy controls, and IGFBP2 appeared to be increased in risk diseases of pancreatic malignancy, such as intraductal papillary mucinous neoplasms (IPMNs). Furthermore, diagnosis of IDACP using the combination of carbohydrate antigen 19–9 (CA19-9), IGFBP2 and IGFBP3 is significantly more effective than CA19-9 alone. This suggests that IGFBP2 and IGFBP3 may serve as compensatory biomarkers for CA19-9. Early diagnosis with this marker combination may improve the prognosis of IDACP patients.

Gathering vast data sets of cancer genomes requires more efficient and autonomous procedures to classify cancer types and to discover a few essential genes to distinguish different cancers. Because protein expression is more stable than gene expression, we chose reverse phase protein array (RPPA) data, a powerful and robust antibody-based high-throughput approach for targeted proteomics, to perform our research. In this study, we proposed a computational framework to classify the patient samples into ten major cancer types based on the RPPA data using the SMO (Sequential minimal optimization) method. A careful feature selection procedure was employed to select 23 important proteins from the total of 187 proteins by mRMR (minimum Redundancy Maximum Relevance Feature Selection) and IFS (Incremental Feature Selection) on the training set. By using the 23 proteins, we successfully classified the ten cancer types with an MCC (Matthews Correlation Coefficient) of 0.904 on the training set, evaluated by 10-fold cross-validation, and an MCC of 0.936 on an independent test set. Further analysis of these 23 proteins was performed. Most of these proteins can present the hallmarks of cancer; Chk2, for example, plays an important role in the proliferation of cancer cells. Our analysis of these 23 proteins lends credence to the importance of these genes as indicators of cancer classification. We also believe our methods and findings may shed light on the discoveries of specific biomarkers of different types of cancers.

Rift Valley fever virus (RVFV) infects both ruminants and humans leading to a wide variance of pathologies dependent on host background and age. Utilizing a targeted reverse phase protein array (RPPA) to define changes in signaling cascades after in vitro infection of human cells with virulent and attenuated RVFV strains, we observed high phosphorylation of Smad transcription factors. This evolutionarily conserved family is phosphorylated by and transduces the activation of TGF-β superfamily receptors. Moreover, we observed that phosphorylation of Smad proteins required active RVFV replication and loss of NSs impaired this activation, further corroborating the RPPA results. Gene promoter analysis of transcripts altered after RVFV infection identified 913 genes that contained a Smad-response element. Functional annotation of these potential Smad-regulated genes clustered in axonal guidance, hepatic fibrosis and cell signaling pathways involved in cellular adhesion/migration, calcium influx, and cytoskeletal reorganization. Furthermore, chromatin immunoprecipitation confirmed the presence of a Smad complex on the interleukin 1 receptor type 2 (IL1R2) promoter, which acts as a decoy receptor for IL-1 activation.

RPPA data analysis

Systematic alanine-scanning of the linear peptide bisebromoamide (BBA), isolated from a marine cyanobacterium, is enabled by targeting the solid phase peptide synthesis of thiazole analogues. The synthetic Tz-BBA analogues have comparable cytotoxicity (nM) to bisebromoamide and cellular morphology assays indicate that they target the actin cytoskeleton. Pathway inhibition in human colon tumour (HCT-116) cells has been explored using reverse phase protein array (RPPA) analysis, which shows a dose-dependent response of IRS-1 expression. Alanine-scanning reveals a structural dependence to the cytotoxicity, actin-targeting and pathway inhibition, and allows a new readily-synthesised lead to be proposed.

Table S1: IC50 data; Table S2: RPPA normalized raw data; Table S3: Differentially expressed analysis in all RNA-seq data; Table S4: Summary of statistical analysis results cell cycle; Table S5: Mutational landscape of the PDX1415 and PDX1526; Table S6: Focal copy number variations of the PDX1415 and PDX1526 models; Table S7: Beta score of the CRISPR/CAS9 KO screening; Table S8: Antibodies

Additional file 2. Number of samples from each TCGA cancer type that are “dropped” as more data types are added to the analysis. The “base” column indicates the number of samples that are present per cancer type in the final intersection of all data types (i.e. each sample counted in the last column has data for each of the 7 data types, including gene expression (not listed here) and somatic mutations).

Enrichment analysis from TCGA RNA-seq and RPPA data comparing CREBBPaltered versus wild-type patients.

Supplementary Table 1. Dataset for the reverse-phase protein array (RPPA) and processed data for Ingenuity Pathway Analysis.

Additional file 8: Table S1. Clinical features and manual pathology assessment of tumor purity throughout the depths of USC patient specimen blocks. Representative H&E stained tissue sections on glass slides were examined at ~100 µm intervals by a board-certified pathologist for evaluation of the relative contributions (as percentages) of tumor cellularity, stroma, necrosis, normal ovarian epithelium, lymphocytes, and polymorphonuclear leukocytes (PMN) to the overall tissue composition. Multiple images per level/case were reviewed; the median estimates of tumor and stroma cellularity are reported with corresponding coefficient of variation (%CV) reported in parentheses. Abbreviations: NACT= neoadjuvant chemotherapy; NOS= not otherwise specified. Table S2. Depiction of study cohort. Numerical values indicate the number of LMD tissue sections that were used for each collection. Greyed boxes represent samples that were not collected or did not have sufficient yield of the target analyte for analysis. Table S3. Global protein matrix. Log2 transformed fold-change abundances of 6503 proteins co-quantified across all samples (n=118). Table S4. Transcriptome matrix for case 343VY. Log2-transformed normalized abundances of 15,558 RNA transcripts measured in case 343VY calculated relative to the average RPM abundance quantified across all samples for a given transcript. Table S5. Log2-transformed target-wise median centered RPPA abundances of protein and phosphoprotein targets in ET, ES, and BT samples. Table S6. Cell type enrichment scores using transcriptomic data for case 343VY in xCell ( http://xcell.ucsf.edu/ , [21]). Table S7. Cell type enrichment scores using proteomic data in xCell ( http://xcell.ucsf.edu/ , [21]). Table S8. ssGSEA scores calculated from “tumor”, “stroma”, and “immune” classifiers. Table S9. Median absolute deviation (MAD) of LMD enriched samples expressing or lacking signal peptide sequences and extracellular classification. P-values indicate the reliability of the presence or absence of a signal peptide or extracellular classification within the indicated LMD enriched tissue across all levels/case. Table S10. Spearman correlations for co-quantified proteins and transcripts from case 343VY. Spearman correlations were calculated using Log2 transformed fold-change abundances of 6,019 imputed proteins that were co-measured as transcripts. Table S11. Spearman correlations between samples using proteins co-quantified by MS and RPPA. Table S12. Spearman correlations for proteins co-quantified by MS and RPPA. Table S13. Pairwise Spearman correlations within and between ET and ES samples using proteins with MAD>1 for construction of patient-specific dendrograms. Table S14. Pairwise Spearman correlations between BT harvests using proteomic abundances. Table S15. LogFC values of proteins measured in HGSOC specimens which were commonly differentially expressed (limma adj. p<0.05) in ET and ES from USC specimens. LogFC protein abundances from Hunt et al Table S7 [13] which displayed the same pattern of expression and passed limma adj. p<0.05 across all patients were prioritized for comparative analysis with LMD enriched samples from USC specimens. The median LogFC values for 313 proteins co-altered from HGSOC samples, which were used as input for Ingenuity Pathway Analysis (IPA). Proteins reported in this table correspond to the HGSOC LogFC values for proteins in the center panel of the venn diagram in Fig. 6. Table S16. LogFC values of proteins measured in USC specimens which were commonly differentially expressed (limma adj. p<0.05) in ET and ES from HGSOC specimens. LogFC protein abundances from USC LMD enriched samples which displayed the same pattern of expression across all patients were prioritized for comparative analysis with LMD enriched samples from HGSOC specimens. The median LogFC values for 313 proteins co-altered from USC samples, which were used as input for Ingenuity Pathway Analysis (IPA). Proteins reported in this table correspond to the USC LogFC values for proteins in the center panel of the venn diagram in Fig. 6. Table S17. LogFC values of proteins measured in HGSOC specimens which were uniquely differentially expressed (limma adj. p<0.05) in ET and ES, which were not co-altered in LMD enriched samples from USC specimens. The 483 proteins reported in this table correspond to the HGSOC only LogFC values for proteins in the left panel of the venn diagram in Fig. 6. Table S18. LogFC values of proteins measured in USC specimens which were uniquely differentially expressed (limma adj. p<0.05) in ET and ES, which were not co-altered in LMD enriched samples from HGSOC specimens. The 142 proteins reported in this table correspond to the USC only LogFC values for proteins in the right panel of the venn diagram in Fig. 6. Table S19. Drug targets and canonical pathways identified by Ingenuity Pathway Analysis (IPA) significantly altered in ET and ES from USC and/or HGSOC specimens. Targets and pathways designated as “HGSOC only” correspond to the those identified using the 483 uniquely differentially expressed proteins between ET and ES from Additional file 8: Table S16 (Fig. 6, left panel of venn diagram). Targets and pathways designated as “USC only” correspond to the those identified using the 142 uniquely differentially expressed between ET and ES from Additional file 8: Table S17 (Fig. 6, right panel of venn diagram). Targets and pathways designated as “Overlap” correspond to those identified when using the HGSOC and USC (as specified) LogFC values of the 313 proteins in Additional file 8: Tables S14 and S15, respectively, as input.

Supplementary Table 5: The tumor cellularity of each biospecimen as assessed by H&E Supplementary Table 6: Correlation of protein expression by RPPA between different biospecimens (Spearman's correlation). Analysis was conducted in samples with 20% cellularity as well as in all samples separately. Supplementary Table 7: Comparison of protein expression by RPPA between different biospecimens (Student's t-test) Supplementary Table 8: Correlation of biomarker expression by RPPA between CNB, central and peripheral samples based on biospecimen variables Supplementary Table 9: Correlation of biomarker expression by RPPA and tumor cellularity in CNB, central and peripheral samples. Analysis was conducted in samples with 20% cellularity as well as in all samples separately. Supplementary Table 10: Correlation of biomarker expression by RPPA in CNB and FNA with central and peripheral samples based on whether neoadjuvant chemotherapy had been administered.

Summary:The datasets described here were gathered while investigating the molecular processes by which some human ductal carcinoma in situ (DCIS) lesions advance to the more aggressive form while others remain indolent.Data access:All RNA sequencing data have been deposited in the Gene Expression Omnibus with accession https://identifiers.org/geo:GSE143790 All Chip-Exo data have been deposited in the Gene Expression Omnibus with accession https://identifiers.org/geo:GSE143313 RPPA data is included together with this data record, in the file Supplementary Figure 3-RPPA.xlsx. The Raw RPPA and ANOVA tabs are the results, while the other tabs are IPA analysis performed by the authors. Permission to use figures and data generated using QIAGEN Ingenuity Pathway Analysis (IPA) is given in the file QIAGEN Ingenuity Product Support Permission letter for Dr. Behbod.pdf.The specific data underlying each figure and supplementary figure in the manuscript are provided as part of this data record, and are as follows:Figure 1-BCL9-STAT3 interaction.xlsxFigure 2-ChIP Exo.xlsxFigure 3-ChIP.xlsxFigure 4-MMP16 and avb3 MIND xenografts.xlsxFigure 5-MMP16 avb3 TMA analysis.xlsxFigure 6-Carnosic data.xlsxSupplementary Figure 3-RPPA.xlsxSupplementary Figure 4-STAT3 Reporter.xlsxSupplementary Figure 5-ChIP Exo Motifs.xlsxSupplementary Figure 6-integrin data.xlsxSupplementary Figure 7-MMP data.xlsxStudy approval and patient consent: Patients gave written informed consent for participation in the University of Kansas Medical Center Institutional Review Board–approved study allowing collection of additional biopsies and or surgical tissue for research. Animal experiments were conducted following protocols approved by the University of Kansas School of Medicine Animal Care and Use and Human Subjects Committee. Study aims and methodology: The aim of the related study was to determine the molecular processes underlying progression to invasion in DCIS using PDX DCIS MIND animal models. Using a novel intraductal model they identify downregulation of specific STAT3 targets to promote progression and use a purified component from rosemary extract to show that treatment in vivo decreases DCIS progression in patient derived DCIS and cell line models.

SummaryThis metadata record provides details of the data supporting the claims of the related manuscript: "Mechanism of action biomarkers predicting response to AKT inhibition in the I-SPY2 breast cancer trial".The related study aimed to determine whether the AKT signaling axis proteins/genes specifically predicted response to AKT inhibitor MK2206 (M). To this end, 26 phospho-proteins and 10 genes involved in AKTmTOR-HER signaling were tested.The data are gene expression and reverse phase protein array (RPPA) data from 150 women with high-risk stage II and III early breast cancer who were enrolled in the multicenter, multi-arm, neo-adjuvant I-SPY 2 TRIAL (NCT01042379; IND105139).Data accessDe-identified molecular and clinical data have been deposited in NCBI's Gene Expression Omnibus. The superseries accession number is GSE150576. The constituent series accession number for the gene expression data is GSE149322, and the constituent series accession number for the RPPA protein/phospho-protein data is GSE150575. In addition, linear transformation parameters (gene expression) and normalisation parameters (mean, sd per RPPA endpoint) to transform raw to normalised data are available as supplemental files on Gene Expression Omnibus, along with the normalised data matrix used in the analysis (gene expression file: GSE149322_ExpDat_ISPY2_MK2206_n150.txt.gz).Name of Institutional Review Board or ethics committee that approved the studyThe protocol used in the study was approved by Institutional Review Boards at all participating institutions: University of California, San Francisco; George Mason University; Quantum Leap Healthcare Collaborative; University of Texas, MD Anderson Cancer Center; Berry Consultants, LLC.

Calculations for predictions of drug response in cell lines, based on RPPA data (used for Challenge submission). (XLSX)