A network-guided penalized regression with application to proteomics data.

Mission: The Clinical Proteomic Tumor Analysis Consortium (CPTAC) is a comprehensive and coordinated effort to accelerate the understanding of the molecular basis of cancer through the application of robust, quantitative, proteomic technologies and workflows. Goals: The overarching goal of CPTAC is to improve our ability to diagnose, treat and prevent cancer. To achieve this goal in a scientifically rigorous manner, the National Cancer Institute (NCI) launched CPTAC to systematically identify proteins that derive from alterations in cancer genomes and related biological processes, and provide this data with accompanying assays and protocols to the public. Genomics initiatives such as The Cancer Genome Atlas (TCGA) have characterized and sequenced the genomic alterations from several types of cancer. CPTAC will leverage its analytical outputs by producing a unique continuum that defines the proteins translated from cancer genomes in order to link genotype to proteotype and ultimately to phenotype. Scientific Approach: The structure of CPTAC is designed to facilitate the efficient implementation of a two-step developmental pipeline, Discovery and Verification. The pipeline originates with the CPTAC Resource Center, which will provide a set of high-quality, genomically characterized clinical biospecimens from patients with a single type of cancer. The Discovery Unit of each Proteome Characterization Center (PCC) will use a set of proteomic technologies to comprehensively analyze the protein composition of this set of biospecimens including mapping proteome to genome and detecting and quantifying protein products that correspond to genomic aberrations. All data from the Discovery Units will be shared within the network via the CPTAC Data Center. Based on this aggregate data set, each Discovery Unit will select and prioritize cancer-related proteins for potential use in the verification step. The selected lists of prioritized proteins by the Discovery Units will be consolidated and collectively prioritized. The Verification Unit of each PCC will develop targeted assays against corresponding protein targets as assigned by the verification plan. The Verification Units will then verify the biomarker candidates by performing these assays in clinically relevant cancer biospecimens (provided by the CPTAC Resource Center). Data from these verification studies will be shared via the CPTAC Data Center. These verified proteins will serve as high-value targets to other initiatives involved in clinical qualification studies. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1282. doi:1538-7445.AM2012-1282

Alzheimer's disease (AD) is a heterogeneous neurodegenerative disease with complex pathophysiology. Therefore, the identification of novel effective fluid biomarkers is essential for Alzheimer's disease diagnosis and drug development. This study aimed to identify potential candidate hub proteins in cerebrospinal fluid for precise Alzheimer's disease diagnosis using bioinformatics methods. A total of 29 co-significant differentially expressed proteins were identified by differential protein expression analysis in four different cohorts. Functional enrichment analysis revealed that most of these proteins were enriched in pathways related to glycometabolism. Using the Least Absolute Shrinkage and Selection Operator (LASSO) and random forest feature selection methods, six hub proteins [14-3-3 protein zeta/delta (YWHAZ), SPARC-related modular calcium-binding protein 1 (SMOC1), aldolase A (ALDOA), pyruvate kinase isoenzyme type M2 (PKM), chitinase-3-like protein 1 (CHI3L1), and secreted phosphoprotein 1 (SPP1)] were identified. These six hub proteins were upregulated in the cerebrospinal fluid of patients with Alzheimer's disease compared with cognitively unimpaired control individuals. Meanwhile, SMOC1, ALDOA, and PKM were specifically upregulated in the cerebrospinal fluid of patients with Alzheimer's disease but not in other neurodegenerative diseases. Build AD diagnostic models showed that a single hub protein or six hub proteins combination had an excellent ability to discriminate Alzheimer's disease. In conclusion, our study suggests that these identified hub proteins, which are related to glycometabolism, may be potential biomarkers for further basic and clinical research in Alzheimer's disease.

PTMcosmos is a comprehensive database with an interactive web portal designed to catalog and visualize post-translational modifications (PTMs) in humans. It contains 469,183 experimentally-validated PTM sites and their supporting evidence from UniProt Knowledge Base, PhosphoSitePlus, and the Clinical Proteomic Tumor Analysis Consortium (CPTAC). PTMcosmos summarizes the entire spectrum of CPTAC proteomics data on human cancer patients, including protein and PTM peptide abundance data from 10 different cancer types. Additionally, PTMcosmos contains cancer somatic mutations from The Cancer Genome Atlas (TCGA), thus allowing for the collective integration and analysis of different data types. In PTMcosmos, we have built an ensemble of interactive visualization tools that allow researchers to investigate altered PTM functions due to genetic alterations in close proximity. The database is live at https://ptmcosmos.wustl.edu. We used PTMcosmos to investigate PTM regulation across cancer types. First, we examined the differential abundance of the PTM sites of cancer driver genes, focusing primarily on phosphorylation of the tumor suppressor retinoblastoma protein (encoded by RB1) and acetylation of the histone acetyltransferase E1A Binding Protein P300 (EP300) across cancer types. We analyzed the association between these PTM events and downstream targets, as well as with tumor subtypes, significantly mutated gene (SMG) mutation status, and clinical features. Second, we investigated the association of the protein abundance of cancer driver genes with ubiquitylsites in lung squamous cell carcinoma (LSCC) to nominate potentially novel modes of regulation of these proteins’ activities. We further analyzed the tumor subtype specificity and tumor-normal abundance changes of these ubiquitylsites and their corresponding substrate proteins, identifying several EGFR ubiquitylsites which may regulate EGFR abundance in LSCC. Finally, we identified the linear and spatial clustering of mutations and PTM sites, identifying multiple mutation-PTM clusters in cancer related genes, including TP53, PIK3CA, CTNNB1, EGFR, and IDH1. We envision that PTMcosmos will serve both the CPTAC consortium and the wider research community to better understand the role of PTMs in cancer. Citation Format: Liang-Bo Wang, Akshay Govindan, Song Cao, Li Ding. PTMcosmos: A web portal of post-translational modifications and proteogenomic resources in cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5027.

BackgroundHepatocellular carcinoma (HCC), is the fifth most common cancer in the world and the second most common cause of cancer-related deaths. Over 500,000 new HCC cases are diagnosed each year. Combining advanced genomic analysis with proteomic characterization not only has great potential in the discovery of useful biomarkers but also drives the development of new diagnostic methods.MethodsThis study obtained proteomic data from Clinical Proteomic Tumor Analysis Consortium (CPTAC) and validated in The Cancer Proteome Atlas (TCPA) and TCGA dataset to identify HCC biomarkers and the dysfunctional of proteogenomics.ResultsThe CPTAC database contained data for 159 patients diagnosed with Hepatitis-B related HCC and 422 differentially expressed proteins (112 upregulated and 310 downregulated proteins). Restricting our analysis to the intersection in survival-related proteins between CPTAC and TCPA database revealed four coverage survival-related proteins including PCNA, MSH6, CDK1, and ASNS.ConclusionThis study established a novel protein signature for HCC prognosis prediction using data retrieved from online databases. However, the signatures need to be verified using independent cohorts and functional experiments.

The Clinical Proteomic Tumor Analysis Consortium (CPTAC), under the auspices of the National Cancer Institute's Office of Cancer Clinical Proteomics Research, is a comprehensive and coordinated effort to accelerate the understanding of the molecular basis of cancer through the application of proteomic technologies and workflows to clinical tumor samples with characterized genomic and transcript profiles. The consortium analyzes cancer biospecimens using mass spectrometry, identifying and quantifying the constituent proteins and characterizing each tumor sample's proteome. Mass spectrometry enables highly specific identification of proteins and their isoforms, accurate relative quantitation of protein abundance in contrasting biospecimens, and localization of post-translational protein modifications, such as phosphorylation, on a protein's sequence. The combination of proteomics, transcriptomics, and genomics data from the same clinical tumor samples provides an unprecedented opportunity for tumor proteogenomics. The CPTAC Data Portal is the centralized data repository for the dissemination of proteomic data collected by Proteome Characterization Centers (PCCs) in the consortium. The portal currently hosts 6.3 TB of data and includes proteomic investigations of breast, colorectal, and ovarian tumor tissues from The Cancer Genome Atlas (TCGA). The data collected by the consortium is made freely available to the public through the data portal.

Pancreatic Ductal Adenocarcinoma (PDAC) is a highly aggressive cancer with 5-year survival rate less than 10%. Although numerous efforts have been made to characterize the functional driver genes of PDAC, the impacts of genomic alterations on protein modifications and molecular mechanism of this cancer type are still not well defined. To this end, Investigators from The Clinical Proteomic Tumor Analysis Consortium (CPTAC) conducted the first comprehensive characterization of 140 pancreatic cancers, 67 normal adjacent tissues, and 9 normal pancreatic ductal tissues using whole genome sequencing, whole exome sequencing, methylation, RNA-seq, miRNA-seq, proteomics, phosphoproteomics, and glycoproteomics. To address the inherent low neoplastic cellularity of pancreatic cancer, multiple orthogonal strategies using molecular features and histology imaging were deployed to deconvolute the cellularity and identify 105 tumors with sufficient neoplastic purity for downstream analyses. Proteomics, phosphoproteomics, and glycoproteomics analyses by mass spectrometry identified and quantified 11,662 proteins, 51,469 phosphosites, and 34,024 glycopeptides, respectively. Genomic data revealed that 97% of the cancers harbored KRAS alterations in our cohort. In addition, we found that somatic mutations and copy number alterations could impact gene and protein expression, as well as protein modifications such as phosphorylation and glycosylation. Glycoproteomic analyses uncovered tumor-associated alterations in glycoprotein biosynthesis and glycoproteins as potential targets for diagnosis or therapeutic intervention. Over-expressed kinase substrates and their corresponding kinases were identified with integration of global proteomic and phosphoproteomic measurements, thus identifying potential therapeutic targets. The results of our molecular and cellular subtyping revealed a small group of immune-hot subtype tumors that could benefit from immunotherapy, as well as the underlying mechanisms associated with major groups of immune-cold subtypes, including endothelial cell remodeling, glycolysis, and cell junction dysregulation. Non-negative matrix factorization (NMF)-based proteogenomics subtyping revealed two clusters with strong prognostic relevance. Finally, we identified proteins and glycoproteins overexpressed in early stage pancreatic cancer that may serve as candidates for early detection. This comprehensive proteogenomic characterization of PDAC provides a valuable resource for uncovering the molecular mechanisms of this cancer and thus paves the way for discovery of novel early detection and therapeutic targets. Citation Format: Liwei Cao, Chen Huang, Daniel Cui Zhou, Oliver F. Bathe, Daniel W. Chan, Ralph H. Hruban, Li Ding, Bing Zhang, Hui Zhang, Clinical Proteomic Tumor Analysis Consortium Investigators. Proteogenomic characterization of pancreatic ductal adenocarcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 17.

The Cancer Genome Atlas project included ovarian serous cystadenocarcinoma as one of the first four tumor types examined, and TCGA has currently completed comprehensive genomic analyses of over 500 HGSOC tumor specimens, including complete exome sequencing, SNP and methylation analysis, as well as mRNA and miRNA profiling. The Clinical Proteomics Tumor Analysis Consortium (CPTAC) has collaborated with TCGA to provide equally comprehensive proteomic analysis of the same tumors analyzed by TCGA, including quantitative global proteomics, phosphoproteomics, and glycoproteomics. The objective of this study is to further understanding of the execution of genome-level aberrations at the protein level, and to identify functional changes associated with clinical outcome that may provide new insights into the mechanisms of disease progression, factors contributing to chemoresistance, and potentially the identification of new therapeutic targets. Initial analysis of a subset of the TCGA ovarian tumors at Pacific Northwest National Laboratory and Johns Hopkins Medical Institutes has provided preliminary evidence of pathway-level dysregulation associated with platinum resistance and poor clinical outcome. Citation Format: Karin D. Rodland. Comprehensive proteomic analyses of TCGA ovarian cancer specimens: Update from CPTAC. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research: From Concept to Clinic; Sep 18-21, 2013; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2013;19(19 Suppl):Abstract nr IA10.

Mass spectrometry (MS) based targeted proteomic methods such as selected reaction monitoring (SRM) are emerging as a promising tool for verification of candidate proteins in biological and biomedical applications. The Clinical Proteomic Tumor Analysis Consortium (CPTAC) of the National Cancer Institute has investigated the standardization and analytical validation of the SRM assays and demonstrated robust analytical performance on different instruments across different laboratories. An Assay Portal has also been established by CPTAC to provide the research community a resource consisting of large sets of targeted MS-based assays, and a depository to share assays publicly. Herein, we report the development of 98 SRM assays that have been thoroughly characterized according to the CPTAC Assay Characterization Guidance Document; 37 of these passed all five experimental tests. The assays cover 70 proteins previously identified at the protein level in ovarian tumors. The experiments, methods and results for characterizing these SRM assays for their MS response, repeatability, selectivity, stability, and endogenous detection are described in detail. Data are available via PeptideAtlas, Panorama and the CPTAC Assay Portal.

The National Cancer Institute's Clinical Proteomic Tumor Analysis Consortium (CPTAC) is an international effort to accelerate the understanding of the molecular basis of cancer through the application of large-scale proteomic and genomic analysis, or proteogenomics. Leidos Biomedical Research, a prime contractor to NCI, provides a robust infrastructure for supporting the collection and sequencing of high-quality biospecimens and data, in addition to project and subcontract management for this program. CPTAC applies the understanding of the molecular basis of cancer to identify biomarker candidates. CPTAC phase III began late 2016 to collect and analyze 200 cases of each of ten additional cancers. The cancers include sarcoma, pancreatic ductal adenocarcinoma, head and neck squamous cell carcinoma, lung adenocarcinoma, lung squamous cell carcinoma, uterine corpus endometrial carcinoma, clear cell renal cell carcinoma, glioblastoma multiforme, and acute myeloid leukemia. The program completed the discovery phase collection of 100 cases for 3 tumor types in the first year. The study entails collection and pathology evaluation of biospecimens, high-quality clinical data and images from clinical sites around the world. A biorepository evaluates and processes the biospecimens, sending nucleic acids to sequencing centers and tissues to proteomics groups. Data are combined and analyzed by proteogenomic data analysis and translational centers. Genomic data are made available to the research community through the NCI Genomic Data Commons. Proteomic and proteogenomic data are made available through the Data Coordinating Center. Imaging data are made available through NCI imaging archive. We report here on progress in collection of tissues and clinical data, the preliminary analysis trends and the integration of biorepository, proteomics and genomics and analysis centers. Citation Format: Mathangi Thiagarajan. NCI CPTAC phase III, proteogenomic analysis of 10 cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 784.

The interaction between RNA binding protein (RBP) and circular RNA (circRNA) is important for the regulation of tumor progression. This study aimed to identify the RBP-circRNA network in hepatocellular carcinoma (HCC). 22 differentially expressed (DE) circRNAs in HCC were screened out from Gene Expression Omnibus (GEO) database and their binding RBPs were predicted by Circular RNA Interactome. Among them, 17 DERBPs, which were commonly dysregulated in HCC from The Clinical Proteomic Tumor Analysis Consortium (CPTAC), The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC) projects, were utilized to construct the RBP-circRNA network. Through survival analysis, we found TARDBP was the only prognostic RBP for HCC in CPTAC, TCGA and ICGC projects. High expression of TARDBP was correlated with high grade, advanced stage and low macrophage infiltration of HCC. Additionally, gene set enrichment analysis showed that dysregulated TARDBP might be involved in some pathways related to the HCC pathogenesis. Therefore, a hub RBP-circRNA network was generated based on TARDBP. RNA immunoprecipitation and RNA pull-down confirmed that hsa_circ_0004913 binds to TARDBP. These findings indicated certain RBP-circRNA regulatory network potentially involved in the pathogenesis of HCC, which provides novel insights into the mechanism study and biomarker identification for HCC.

BackgroundNumerous cancer types present the aberrant TANK-binding kinase 1 (TBK1) expression, which plays an important role in driving inflammation and innate immunity. However, the prognostic role of TBK1 and its relationship with immune cell infiltration in hepatocellular carcinoma (HCC) remain unclear.MethodsThe expression and prognostic value of TBK1 was analyzed by Tumor Immune Estimation Resource (TIMER), Kaplan-Meier plotter and Gene Expression Profiling Interactive Analysis (GEPIA), Clinical Proteomic Tumor Analysis Consortium (CPTAC) and further confirmed in the present cohort of patients with HCC. The association between TBK1 and HCC immune infiltrates, and its potential mechanism were investigated via analyses of the Tumor Immune Estimation Resource, tumor-immune system interactions database (TISIDB), CIBERSORT, STRING, and Metascape. The effect of TBK1 on immune infiltrates and the therapeutic value of targeting TBK1 were further investigated in a HCC mouse model by treatment with a TBK1 antagonist.ResultsThe level of TBK1 expression in HCC was higher than that measured in normal tissues, and associated with poorer overall survival (GEPIA: hazard ratio [HR]=1.80, P=0.038; Kaplan–Meier plotter: HR=1.87, P<0.001; CPTAC: HR=2.23, P=0.007; Our cohort: HR=2.92, P=0.002). In addition, high TBK1 expression was found in HCC with advanced TNM stage and identified as an independent poor prognostic factor for overall survival among patients with HCC. In terms of immune infiltration, tumor tissues from HCC patients with high TBK1 expression had a low proportion of CD8+ T cells, and TBK1 expression did not show prognostic value in HCC patients with enriched CD8+ T cells. Furthermore, TBK1 expression was positively correlated with the markers of T cell exhaustion and immunosuppressive cells in the HCC microenvironment. Mechanistically, the promotion of HCC immunosuppression by TBK1 was involved in the regulation of inflammatory cytokines. In vivo experiments revealed that treatment with a TBK1 antagonist delayed HCC growth by increasing the number of tumor-infiltrating CD8+ T cells.ConclusionsThe up-regulated expression of TBK1 may be useful in predicting poor prognosis of patients with HCC. In addition, TBK1, which promotes the HCC immunosuppressive microenvironment, may be a potential immunotherapeutic target for patients with HCC.

Background: Determining a patient’s response to a specific therapy is a vital step in developing personalized cancer treatment. Personalized treatment relies on two key technological advancements: algorithms tailored to each patient’s molecular profiles, and comprehensive datasets detailing the effects of patient diversity on drug response. Despite recent advancements in machine learning-based models, each new algorithm requires curation of numerous datasets, which poses a significant burden to researchers who need to evaluate and compare various algorithms. Currently, datasets include extensive genomics, transcriptomics, proteomics and metabolomics measurements along with drug response data (i.e., cell viability) for each cell line, collected through the Cancer Cell Line Encyclopedia (CCLE) and other initiatives. Additional omics data is available for tumor organoids and patient samples generated by consortia such as Human Cancer Model Initiative (HCMI) and the Clinical Proteomic Tumor Analysis Consortium (CPTAC). However, to date there has been little public effort to annotate and harmonize these data across repositories. Methods: To address this need, we developed the Python package coderdata that collates the most up-to-date cancer omics and drug response data. We curated four public cancer datasets, establishing reproducible pipelines that standardize data retrieval, formatting, and integration so the datasets can be built on a cluster if needed. The datasets undergo bi-monthly synchronization with FigShare, ensuring the data is up to date and available. Scientists can use coderdata to retrieve and format data on the command line interface as well as within Python scripts. Results: Coderdata is a straightforward Python package that will enable the development and benchmarking of diverse machine learning applications in cancer research. At its core, it represents reproducible data harmonization that enables heterogeneous datasets to be analyzed in bulk. This Python package downloads and standardizes data from multiple consortia-related resources including HCMI, CPTAC, CCLE, and BeatAML, together representing 4931 multi-omics samples from across 250+ cancer types. These datasets include cancer data metrics such as copy number, mutation, transcriptomics, proteomics, miRNA, methylation, and drug response data. Conclusions: The applications of machine learning in cancer biology are hampered by the distributed nature of existing datasets. As such, the collection and standardization of data by coderdata substantially reduces the time investment required for data curation in cancer research. Direct access to benchmark multi-omics and drug response datasets enables scientists to focus on algorithm development for tumor drug response prediction, potentially accelerating the discovery of therapeutic strategies. Citation Format: Jeremy Jacobson, Sydney Schwartz, M. Ryan Weil, Neeraj Kumar, Sara Gosline. The cancer omics and drug experimental response dataset (CODERData): A harmonized benchmark dataset for machine learning models of drug response prediction [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6210.

Acute myeloid leukemia (AML) is a heterogeneous cancer of the blood and bone marrow with poor prognosis. To elucidate the underlying mechanisms that drive AML as part of the Clinical Proteomic Tumor Analysis Consortium (CPTAC) effort, we performed comprehensive genomics, transcriptomics, proteomics including multiple post-translational modifications (PTMs), metabolomics, and lipidomics characterization of 173 treatment-naïve AML patients. By applying similarity network fusion using collected transcriptomics and proteomics data jointly, we identified 8 proteogenomic subtypes. These subtypes not only harbored recurrent genomic aberrations of AML similar to the current clinical classification but also unveiled new subtype-specific phenotypes. Utilizing previously published single-cell RNAseq data as reference, we also demonstrated that these subtypes are also enriched for cellular hierarchies including primitive AML, CD14+ monocyte-like, and GMP-like AML. Interestingly, proteogenomic subtyping revealed a distinct NPM1-mutant subset of AMLs characterized by outlier transcription factor expression and activities of FOXC1 and POU2F1. Further, we derived a shared primitive signature across NPM1 mutated AML and myelodysplastic syndrome related AML that could be beneficial to determine drug response. Among the proposed primitive signatures, we found a protein marker MAP1A was able to predict response of venetoclax and NF-kB activation inhibitor across NPM1 mutated AML and myelodysplastic syndrome related AML. To translate these proteogenomics subtypes into potential clinical applications, we further employed methods to evaluate drug responses of proteogenomics subtypes by integrating CPTAC cohort with the previously published Beat AML cohort, which focuses on ex-vivo drug testing, and revealed subtype-specific drug responses. Together with findings in subtypes with PTM and metabolism aberrations, our study provided new valuable insights and clinical relevance of these newly identified in AML. Citation Format: Yi Hsiao, Alec Chu, Yamei Deng, Chenwei Wang, Jennifer Kyle, Yongchao Dou, Jame Pino, Camilo Posso, Leanne Henry, Lijun Chen, Ginny Xiaohe Li, Mamie Lih, Yifat Geffen, Fengchao Yu, Li Ding, Gilbert S. Omenn, Chandan Kumar, Saravana M. Dhanasekaran, Elie Traer, Jefferey W. Tyner, Hui Zhang, Tao Liu, Sara S. Gosline, Bing Zhang, Arul M. Chinnaiyan, Alexey I. Nesvizhskii, Marcin Cieslik, The CPTAC Consortium. Multi-Omic Characterization of Treatment-Naive Acute Myeloid Leukemia Reveals Subtype-Specific Features and Biomarkers of Drug Response [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3749.

Ovarian cancer is one of the most common gynecological malignancies, ranking third after cervical and uterine cancer. High-grade serous ovarian cancer (HGSOC) is one of the most aggressive subtype, and the late onset of its symptoms leads in most cases to an unfavourable prognosis. Current predictive algorithms used to estimate the risk of having Ovarian Cancer fail to provide sufficient sensitivity and specificity to be used widely in clinical practice. The use of additional biomarkers or parameters such as age or menopausal status to overcome these issues showed only weak improvements. It is necessary to identify novel molecular signatures and the development of new predictive algorithms able to support the diagnosis of HGSOC, and at the same time, deepen the understanding of this elusive disease, with the final goal of improving patient survival. Here, we apply a Machine Learning-based pipeline to an open-source HGSOC Proteomic dataset to develop a decision support system (DSS) that displayed high discerning ability on a dataset of HGSOC biopsies. The proposed DSS consists of a double-step feature selection and a decision tree, with the resulting output consisting of a combination of three highly discriminating proteins: TOP1, PDIA4, and OGN, that could be of interest for further clinical and experimental validation. Furthermore, we took advantage of the ranked list of proteins generated during the feature selection steps to perform a pathway analysis to provide a snapshot of the main deregulated pathways of HGSOC. The datasets used for this study are available in the Clinical Proteomic Tumor Analysis Consortium (CPTAC) data portal (https://cptac-data-portal.georgetown.edu/).

ATPase family AAA domain-containing protein 2 (ATAD2) is significantly up-regulated in many cancer types and contributes to poor patient outcomes. ATAD2 exhibits a multidomain architecture comprising an N-terminal acidic domain, two AAA+ ATPase domains, a bromodomain, and a C-terminal domain. The AAA+ ATPase domain facilitates protein oligomerization and ATP binding, while the bromodomain recognizes acetylated lysine in histones and nonhistone proteins. ATAD2 involvement in cancer extends across multiple signaling pathways, such as Rb-E2F1, PI3K/AKT, and TGF-β1/Smad3, which promotes cell proliferation and cancer progression. Herein, we report that ATAD2 directly interacts with TWIST1, and both N-terminal regions of proteins mediate the interaction. Immunofluorescence experiments suggested that ATAD2 and TWIST1 primarily colocalize in the nucleus. Notably, our qPCR results revealed the functional significance of ATAD2-TWIST1 interaction by demonstrating their synergistic effect on the transcriptional activation of MYC in colorectal carcinoma cell lines. Moreover, the ChIP-qPCR result further indicates that ATAD2 and TWIST1 significantly localize in the promoter of the MYC gene. In addition, analysis of The Cancer Genome Atlas (TCGA) and Clinical Proteomic Tumor Analysis Consortium (CPTAC) data suggests a correlation between ATAD2, TWIST1, and MYC overexpression and poor survival rates in colorectal carcinoma. Lastly, the overexpression of ATAD2 and TWIST1 enhances cell proliferation, emphasizing their role in colorectal carcinoma progression through MYC activation. Together, these results suggest that ATAD2 is a crucial factor in TWIST1-dependent MYC gene activation, resulting in an active ATAD2-TWIST1-MYC axis that contributes to colon cancer cell proliferation.