DrugMap: A quantitative pan-cancer analysis of cysteine ligandability.

Cysteine-focused chemical proteomic platforms have accelerated the clinical development of covalent inhibitors for a wide range of targets in cancer. However, how different oncogenic contexts influence cysteine targeting remains unknown. To address this question, we have developed "DrugMap," an atlas of cysteine ligandability compiled across 416 cancer cell lines. We unexpectedly find that cysteine ligandability varies across cancer cell lines, and we attribute this to differences in cellular redox states, protein conformational changes, and genetic mutations. Leveraging these findings, we identify actionable cysteines in NF-κB1 and SOX10 and develop corresponding covalent ligands that block the activity of these transcription factors. We demonstrate that the NF-κB1 probe blocks DNA binding, whereas the SOX10 ligand increases SOX10-SOX10 interactions and disrupts melanoma transcriptional signaling. Our findings reveal heterogeneity in cysteine ligandability across cancers, pinpoint cell-intrinsic features driving cysteine targeting, and illustrate the use of covalent probes to disrupt oncogenic transcription-factor activity.

DrugMap: A quantitative pan-cancer analysis of cysteine ligandability.

Cysteine-focused chemical proteomic platforms have accelerated the clinical development of covalent inhibitors of a wide-range of targets in cancer. However, how different oncogenic contexts influence cysteine targeting remains unknown. To address this question, we have developed DrugMap , an atlas of cysteine ligandability compiled across 416 cancer cell lines. We unexpectedly find that cysteine ligandability varies across cancer cell lines, and we attribute this to differences in cellular redox states, protein conformational changes, and genetic mutations. Leveraging these findings, we identify actionable cysteines in NFκB1 and SOX10 and develop corresponding covalent ligands that block the activity of these transcription factors. We demonstrate that the NFκB1 probe blocks DNA binding, whereas the SOX10 ligand increases SOX10-SOX10 interactions and disrupts melanoma transcriptional signaling. Our findings reveal heterogeneity in cysteine ligandability across cancers, pinpoint cell-intrinsic features driving cysteine targeting, and illustrate the use of covalent probes to disrupt oncogenic transcription factor activity.

Quantitative immunopeptidomics reveals a tumor stroma-specific target for T cell therapy.

T cell receptor (TCR)-based immunotherapy has emerged as a promising therapeutic approach for the treatment of patients with solid cancers. Identifying peptide-human leukocyte antigen (pHLA) complexes highly presented on tumors and rarely expressed on healthy tissue in combination with high-affinity TCRs that when introduced into T cells can redirect T cells to eliminate tumor but not healthy tissue is a key requirement for safe and efficacious TCR-based therapies. To discover promising shared tumor antigens that could be targeted via TCR-based adoptive T cell therapy, we employed population-scale immunopeptidomics using quantitative mass spectrometry across ~1500 tumor and normal tissue samples. We identified an HLA-A*02:01-restricted pan-cancer epitope within the collagen type VI α-3 (COL6A3) gene that is highly presented on tumor stroma across multiple solid cancers due to a tumor-specific alternative splicing event that rarely occurs outside the tumor microenvironment. T cells expressing natural COL6A3-specific TCRs demonstrated only modest activity against cells presenting high copy numbers of COL6A3 pHLAs. One of these TCRs was affinity-enhanced, enabling transduced T cells to specifically eliminate tumors in vivo that expressed similar copy numbers of pHLAs as primary tumor specimens. The enhanced TCR variants exhibited a favorable safety profile with no detectable off-target reactivity, paving the way to initiate clinical trials using COL6A3-specific TCRs to target an array of solid tumors.

Pitfalls in HLA Ligandomics-How to Catch a Li(e)gand.

Knowledge about the peptide repertoire presented by human leukocyte antigens (HLA) holds the key to unlock target-specific cancer immunotherapies such as adoptive cell therapies or bispecific T cell engaging receptors. Therefore, comprehensive and accurate characterization of HLA peptidomes by mass spectrometry (immunopeptidomics) across tissues and disease states is essential. With growing numbers of immunopeptidomics datasets and the scope of peptide identification strategies reaching beyond the canonical proteome, the likelihood for erroneous peptide identification as well as false annotation of HLA-independent peptides as HLA ligands is increasing. Such "fake ligands" can lead to selection of nonexistent targets for immunotherapeutic development and need to be recognized as such as early as possible in the preclinical pipeline. Here we present computational and experimental methods that enable the identification of "fake ligands" that might be introduced at different steps of the immunopeptidomics workflow. The statistics presented herein allow discrimination of true HLA ligands from coisolated HLA-independent proteolytic fragments. In addition, we describe necessary steps to ensure system suitability of the chromatographic system. Furthermore, we illustrate an algorithm for detection of source fragmentation events that are introduced by electrospray ionization during mass spectrometry. For confirmation of peptide sequences, we present an experimental pipeline that enables high-throughput sequence verification through similarity of fragmentation pattern and coelution of synthetic isotope-labeled internal standards. Based on these methods, we show the overall high quality of existing datasets but point out limitations and pitfalls critical for individual peptides and how they can be uncovered in order to identify true ligands.

Tumor-Selective Altered Glycosylation and Functional Attenuation of CD73 in Human Hepatocellular Carcinoma.

CD73, a cell-surface N-linked glycoprotein that produces extracellular adenosine, is a novel target for cancer immunotherapy. Although anti-CD73 antibodies have entered clinical development, CD73 has both protumor and antitumor functions, depending on the target cell and tumor type. The aim of this study was to characterize CD73 regulation in human hepatocellular carcinoma (HCC). We examined CD73 expression, localization, and activity using molecular, biochemical, and cellular analyses on primary HCC surgical specimens, coupled with mechanistic studies in HCC cells. We analyzed CD73 glycan signatures and global alterations in transcripts encoding other N-linked glycoproteins by using mass spectrometry glycomics and RNA sequencing (RNAseq), respectively. CD73 was expressed on tumor hepatocytes where it exhibited abnormal N-linked glycosylation, independent of HCC etiology, tumor stage, or fibrosis presence. Aberrant glycosylation of tumor-associated CD73 resulted in a 3-fold decrease in 5'-nucleotidase activity (P < 0.0001). Biochemically, tumor-associated CD73 was deficient in hybrid and complex glycans specifically on residues N311 and N333 located in the C-terminal catalytic domain. Blocking N311/N333 glycosylation by site-directed mutagenesis produced CD73 with significantly decreased 5'-nucleotidase activity in vitro, similar to the primary tumors. Glycosylation-deficient CD73 partially colocalized with the Golgi structural protein GM130, which was strongly induced in HCC tumors. RNAseq analysis further revealed that N-linked glycoprotein-encoding genes represented the largest category of differentially expressed genes between HCC tumor and adjacent tissue. Conclusion: We provide the first detailed characterization of CD73 glycosylation in normal and tumor tissue, revealing a novel mechanism that leads to the functional suppression of CD73 in human HCC tumor cells. The present findings have translational implications for therapeutic candidate antibodies targeting cell-surface CD73 in solid tumors and small-molecule adenosine receptor agonists that are in clinical development for HCC.

The HLA ligandome landscape of chronic myeloid leukemia delineates novel T-cell epitopes for immunotherapy.

Antileukemia immunity plays an important role in disease control and maintenance of tyrosine kinase inhibitor (TKI)-free remission in chronic myeloid leukemia (CML). Thus, antigen-specific immunotherapy holds promise for strengthening immune control in CML but requires the identification of CML-associated targets. In this study, we used a mass spectrometry-based approach to identify naturally presented HLA class I- and class II-restricted peptides in primary CML samples. Comparative HLA ligandome profiling using a comprehensive dataset of different hematological benign specimens and samples from CML patients in deep molecular remission delineated a panel of novel frequently presented CML-exclusive peptides. These nonmutated target antigens are of particular relevance because our extensive data-mining approach suggests the absence of naturally presented BCR-ABL- and ABL-BCR-derived HLA-restricted peptides and the lack of frequent tumor-exclusive presentation of known cancer/testis and leukemia-associated antigens. Functional characterization revealed spontaneous T-cell responses against the newly identified CML-associated peptides in CML patient samples and their ability to induce multifunctional and cytotoxic antigen-specific T cells de novo in samples from healthy volunteers and CML patients. Thus, these antigens are prime candidates for T-cell-based immunotherapeutic approaches that may prolong TKI-free survival and even mediate cure of CML patients.

RNA editing derived epitopes function as cancer antigens to elicit immune responses.

In addition to genomic mutations, RNA editing is another major mechanism creating sequence variations in proteins by introducing nucleotide changes in mRNA sequences. Deregulated RNA editing contributes to different types of human diseases, including cancers. Here we report that peptides generated as a consequence of RNA editing are indeed naturally presented by human leukocyte antigen (HLA) molecules. We provide evidence that effector CD8+ T cells specific for edited peptides derived from cyclin I are present in human tumours and attack tumour cells that are presenting these epitopes. We show that subpopulations of cancer patients have increased peptide levels and that levels of edited RNA correlate with peptide copy numbers. These findings demonstrate that RNA editing extends the classes of HLA presented self-antigens and that these antigens can be recognised by the immune system.

Functional proteogenomics reveals biomarkers and therapeutic targets in lymphomas.

Identification of biomarkers and therapeutic targets is a critical goal of precision medicine. N-glycoproteins are a particularly attractive class of proteins that constitute potential cancer biomarkers and therapeutic targets for small molecules, antibodies, and cellular therapies. Using mass spectrometry (MS), we generated a compendium of 1,091 N-glycoproteins (from 40 human primary lymphomas and cell lines). Hierarchical clustering revealed distinct subtype signatures that included several subtype-specific biomarkers. Orthogonal immunological studies in 671 primary lymphoma tissue biopsies and 32 lymphoma-derived cell lines corroborated MS data. In anaplastic lymphoma kinase-positive (ALK+) anaplastic large cell lymphoma (ALCL), integration of N-glycoproteomics and transcriptome sequencing revealed an ALK-regulated cytokine/receptor signaling network, including vulnerabilities corroborated by a genome-wide clustered regularly interspaced short palindromic screen. Functional targeting of IL-31 receptor ß, an ALCL-enriched and ALK-regulated N-glycoprotein in this network, abrogated ALK+ALCL growth in vitro and in vivo. Our results highlight the utility of functional proteogenomic approaches for discovery of cancer biomarkers and therapeutic targets.

New targeted approaches for the quantification of data-independent acquisition mass spectrometry.

The use of data-independent acquisition (DIA) approaches for the reproducible and precise quantification of complex protein samples has increased in the last years. The protein information arising from DIA analysis is stored in digital protein maps (DIA maps) that can be interrogated in a targeted way by using ad hoc or publically available peptide spectral libraries generated on the same sample species as for the generation of the DIA maps. The restricted availability of certain difficult-to-obtain human tissues (i.e., brain) together with the caveats of using spectral libraries generated under variable experimental conditions limits the potential of DIA. Therefore, DIA workflows would benefit from high-quality and extended spectral libraries that could be generated without the need of using valuable samples for library production. We describe here two new targeted approaches, using either classical data-dependent acquisition repositories (not specifically built for DIA) or ad hoc mouse spectral libraries, which enable the profiling of human brain DIA data set. The comparison of our results to both the most extended publically available human spectral library and to a state-of-the-art untargeted method supports the use of these new strategies to improve future DIA profiling efforts.

A multicenter study benchmarks software tools for label-free proteome quantification.

Consistent and accurate quantification of proteins by mass spectrometry (MS)-based proteomics depends on the performance of instruments, acquisition methods and data analysis software. In collaboration with the software developers, we evaluated OpenSWATH, SWATH 2.0, Skyline, Spectronaut and DIA-Umpire, five of the most widely used software methods for processing data from sequential window acquisition of all theoretical fragment-ion spectra (SWATH)-MS, which uses data-independent acquisition (DIA) for label-free protein quantification. We analyzed high-complexity test data sets from hybrid proteome samples of defined quantitative composition acquired on two different MS instruments using different SWATH isolation-window setups. For consistent evaluation, we developed LFQbench, an R package, to calculate metrics of precision and accuracy in label-free quantitative MS and report the identification performance, robustness and specificity of each software tool. Our reference data sets enabled developers to improve their software tools. After optimization, all tools provided highly convergent identification and reliable quantification performance, underscoring their robustness for label-free quantitative proteomics.

Untargeted, spectral library-free analysis of data-independent acquisition proteomics data generated using Orbitrap mass spectrometers.

We describe an improved version of the data-independent acquisition (DIA) computational analysis tool DIA-Umpire, and show that it enables highly sensitive, untargeted, and direct (spectral library-free) analysis of DIA data obtained using the Orbitrap family of mass spectrometers. DIA-Umpire v2 implements an improved feature detection algorithm with two additional filters based on the isotope pattern and fractional peptide mass analysis. The targeted re-extraction step of DIA-Umpire is updated with an improved scoring function and a more robust, semiparametric mixture modeling of the resulting scores for computing posterior probabilities of correct peptide identification in a targeted setting. Using two publicly available Q Exactive DIA datasets generated using HEK-293 cells and human liver microtissues, we demonstrate that DIA-Umpire can identify similar number of peptide ions, but with better identification reproducibility between replicates and samples, as with conventional data-dependent acquisition. We further demonstrate the utility of DIA-Umpire using a series of Orbitrap Fusion DIA experiments with HeLa cell lysates profiled using conventional data-dependent acquisition and using DIA with different isolation window widths.

Data Independent Acquisition analysis in ProHits 4.0.

Affinity purification coupled with mass spectrometry (AP-MS) is a powerful technique for the identification and quantification of physical interactions. AP-MS requires careful experimental design, appropriate control selection and quantitative workflows to successfully identify bona fide interactors amongst a large background of contaminants. We previously introduced ProHits, a Laboratory Information Management System for interaction proteomics, which tracks all samples in a mass spectrometry facility, initiates database searches and provides visualization tools for spectral counting-based AP-MS approaches. More recently, we implemented Significance Analysis of INTeractome (SAINT) within ProHits to provide scoring of interactions based on spectral counts. Here, we provide an update to ProHits to support Data Independent Acquisition (DIA) with identification software (DIA-Umpire and MSPLIT-DIA), quantification tools (through DIA-Umpire, or externally via targeted extraction), and assessment of quantitative enrichment (through mapDIA) and scoring of interactions (through SAINT-intensity). With additional improvements, notably support of the iProphet pipeline, facilitated deposition into ProteomeXchange repositories and enhanced export and viewing functions, ProHits 4.0 offers a comprehensive suite of tools to facilitate affinity proteomics studies. SIGNIFICANCE: It remains challenging to score, annotate and analyze proteomics data in a transparent manner. ProHits was previously introduced as a LIMS to enable storing, tracking and analysis of standard AP-MS data. In this revised version, we expand ProHits to include integration with a number of identification and quantification tools based on Data-Independent Acquisition (DIA). ProHits 4.0 also facilitates data deposition into public repositories, and the transfer of data to new visualization tools.

mapDIA: Preprocessing and statistical analysis of quantitative proteomics data from data independent acquisition mass spectrometry.

UNLABELLED: Data independent acquisition (DIA) mass spectrometry is an emerging technique that offers more complete detection and quantification of peptides and proteins across multiple samples. DIA allows fragment-level quantification, which can be considered as repeated measurements of the abundance of the corresponding peptides and proteins in the downstream statistical analysis. However, few statistical approaches are available for aggregating these complex fragment-level data into peptide- or protein-level statistical summaries. In this work, we describe a software package, mapDIA, for statistical analysis of differential protein expression using DIA fragment-level intensities. The workflow consists of three major steps: intensity normalization, peptide/fragment selection, and statistical analysis. First, mapDIA offers normalization of fragment-level intensities by total intensity sums as well as a novel alternative normalization by local intensity sums in retention time space. Second, mapDIA removes outlier observations and selects peptides/fragments that preserve the major quantitative patterns across all samples for each protein. Last, using the selected fragments and peptides, mapDIA performs model-based statistical significance analysis of protein-level differential expression between specified groups of samples. Using a comprehensive set of simulation datasets, we show that mapDIA detects differentially expressed proteins with accurate control of the false discovery rates. We also describe the analysis procedure in detail using two recently published DIA datasets generated for 14-3-3ß dynamic interaction network and prostate cancer glycoproteome. AVAILABILITY: The software was written in C++ language and the source code is available for free through SourceForge website http://sourceforge.net/projects/mapdia/.This article is part of a Special Issue entitled: Computational Proteomics.

QPROT: Statistical method for testing differential expression using protein-level intensity data in label-free quantitative proteomics.

We introduce QPROT, a statistical framework and computational tool for differential protein expression analysis using protein intensity data. QPROT is an extension of the QSPEC suite, originally developed for spectral count data, adapted for the analysis using continuously measured protein-level intensity data. QPROT offers a new intensity normalization procedure and model-based differential expression analysis, both of which account for missing data. Determination of differential expression of each protein is based on the standardized Z-statistic based on the posterior distribution of the log fold change parameter, guided by the false discovery rate estimated by a well-known Empirical Bayes method. We evaluated the classification performance of QPROT using the quantification calibration data from the clinical proteomic technology assessment for cancer (CPTAC) study and a recently published Escherichia coli benchmark dataset, with evaluation of FDR accuracy in the latter. BIOLOGICAL SIGNIFICANCE: QPROT is a statistical framework with computational software tool for comparative quantitative proteomics analysis. It features various extensions of QSPEC method originally built for spectral count data analysis, including probabilistic treatment of missing values in protein intensity data. With the increasing popularity of label-free quantitative proteomics data, the proposed method and accompanying software suite will be immediately useful for many proteomics laboratories. This article is part of a Special Issue entitled: Computational Proteomics.

Mining Missing Membrane Proteins by High-pH Reverse-Phase StageTip Fractionation and Multiple Reaction Monitoring Mass Spectrometry.

Despite significant efforts in the past decade toward complete mapping of the human proteome, 3564 proteins (neXtProt, 09-2014) are still "missing proteins". Over one-third of these missing proteins are annotated as membrane proteins, owing to their relatively challenging accessibility with standard shotgun proteomics. Using nonsmall cell lung cancer (NSCLC) as a model study, we aim to mine missing proteins from disease-associated membrane proteome, which may be still largely under-represented. To increase identification coverage, we employed Hp-RP StageTip prefractionation of membrane-enriched samples from 11 NSCLC cell lines. Analysis of membrane samples from 20 pairs of tumor and adjacent normal lung tissue was incorporated to include physiologically expressed membrane proteins. Using multiple search engines (X!Tandem, Comet, and Mascot) and stringent evaluation of FDR (MAYU and PeptideShaker), we identified 7702 proteins (66% membrane proteins) and 178 missing proteins (74 membrane proteins) with PSM-, peptide-, and protein-level FDR of 1%. Through multiple reaction monitoring using synthetic peptides, we provided additional evidence of eight missing proteins including seven with transmembrane helix domains. This study demonstrates that mining missing proteins focused on cancer membrane subproteome can greatly contribute to map the whole human proteome. All data were deposited into ProteomeXchange with the identifier PXD002224.

Large-scale determination of absolute phosphorylation stoichiometries in human cells by motif-targeting quantitative proteomics.

Our ability to model the dynamics of signal transduction networks will depend on accurate methods to quantify levels of protein phosphorylation on a global scale. Here we describe a motif-targeting quantitation method for phosphorylation stoichiometry typing. Proteome-wide phosphorylation stoichiometry can be obtained by a simple phosphoproteomic workflow integrating dephosphorylation and isotope tagging with enzymatic kinase reaction. Proof-of-concept experiments using CK2-, MAPK- and EGFR-targeting assays in lung cancer cells demonstrate the advantage of kinase-targeted complexity reduction, resulting in deeper phosphoproteome quantification. We measure the phosphorylation stoichiometry of >1,000 phosphorylation sites including 366 low-abundance tyrosine phosphorylation sites, with high reproducibility and using small sample sizes. Comparing drug-resistant and sensitive lung cancer cells, we reveal that post-translational phosphorylation changes are significantly more dramatic than those at the protein and messenger RNA levels, and suggest potential drug targets within the kinase-substrate network associated with acquired drug resistance.

DIA-Umpire: comprehensive computational framework for data-independent acquisition proteomics.

As a result of recent improvements in mass spectrometry (MS), there is increased interest in data-independent acquisition (DIA) strategies in which all peptides are systematically fragmented using wide mass-isolation windows ('multiplex fragmentation'). DIA-Umpire (http://diaumpire.sourceforge.net/), a comprehensive computational workflow and open-source software for DIA data, detects precursor and fragment chromatographic features and assembles them into pseudo-tandem MS spectra. These spectra can be identified with conventional database-searching and protein-inference tools, allowing sensitive, untargeted analysis of DIA data without the need for a spectral library. Quantification is done with both precursor- and fragment-ion intensities. Furthermore, DIA-Umpire enables targeted extraction of quantitative information based on peptides initially identified in only a subset of the samples, resulting in more consistent quantification across multiple samples. We demonstrated the performance of the method with control samples of varying complexity and publicly available glycoproteomics and affinity purification-MS data.

Label-free quantitative proteomics of CD133-positive liver cancer stem cells.

BACKGROUND: CD133-positive liver cancer stem cells, which are characterized by their resistance to conventional chemotherapy and their tumor initiation ability at limited dilutions, have been recognized as a critical target in liver cancer therapeutics. In the current work, we developed a label-free quantitative method to investigate the proteome of CD133-positive liver cancer stem cells for the purpose of identifying unique biomarkers that can be utilized for targeting liver cancer stem cells. Label-free quantitation was performed in combination with ID-based Elution time Alignment by Linear regression Quantitation (IDEAL-Q) and MaxQuant. RESULTS: Initially, IDEAL-Q analysis revealed that 151 proteins were differentially expressed in the CD133-positive hepatoma cells when compared with CD133-negative cells. We then analyzed these 151 differentially expressed proteins by MaxQuant software and identified 10 significantly up-regulated proteins. The results were further validated by RT-PCR, western blot, flow cytometry or immunofluorescent staining which revealed that prominin-1, annexin A1, annexin A3, transgelin, creatine kinase B, vimentin, and EpCAM were indeed highly expressed in the CD133-positive hepatoma cells. CONCLUSIONS: These findings confirmed that mass spectrometry-based label-free quantitative proteomics can be used to gain insights into liver cancer stem cells.

Label-free quantitative proteomics and N-glycoproteomics analysis of KRAS-activated human bronchial epithelial cells.

Mutational activation of KRAS promotes various malignancies, including lung adenocarcinoma. Knowledge of the molecular targets mediating the downstream effects of activated KRAS is limited. Here, we provide the KRAS target proteins and N-glycoproteins using human bronchial epithelial cells with and without the expression of activated KRAS (KRAS(V12)). Using an OFFGEL peptide fractionation and hydrazide method combined with subsequent LTQ-Orbitrap analysis, we identified 5713 proteins and 608 N-glycosites on 317 proteins in human bronchial epithelial cells. Label-free quantitation of 3058 proteins (≥2 peptides; coefficient of variation (CV) ≤ 20%) and 297 N-glycoproteins (CV ≤ 20%) revealed the differential regulation of 23 proteins and 14 N-glycoproteins caused by activated KRAS, including 84% novel ones. An informatics-assisted IPA-Biomarker® filter analysis prioritized some of the differentially regulated proteins (ALDH3A1, CA2, CTSD, DST, EPHA2, and VIM) and N-glycoproteins (ALCAM, ITGA3, and TIMP-1) as cancer biomarkers. Further, integrated in silico analysis of microarray repository data of lung adenocarcinoma clinical samples and cell lines containing KRAS mutations showed positive mRNA fold changes (p < 0.05) for 61% of the KRAS-regulated proteins, including biomarker proteins, CA2 and CTSD. The most significant discovery of the integrated validation is the down-regulation of FABP5 and PDCD4. A few validated proteins, including tumor suppressor PDCD4, were further confirmed as KRAS targets by shRNA-based knockdown experiments. Finally, the studies on KRAS-regulated N-glycoproteins revealed structural alterations in the core N-glycans of SEMA4B in KRAS-activated human bronchial epithelial cells and functional role of N-glycosylation of TIMP-1 in the regulation of lung adenocarcinoma A549 cell invasion. Together, our study represents the largest proteome and N-glycoproteome data sets for HBECs, which we used to identify several novel potential targets of activated KRAS that may provide insights into KRAS-induced adenocarcinoma and have implications for both lung cancer therapy and diagnosis.

An informatics-assisted label-free approach for personalized tissue membrane proteomics: case study on colorectal cancer.

We developed a multiplexed label-free quantification strategy, which integrates an efficient gel-assisted digestion protocol, high-performance liquid chromatography tandem MS analysis, and a bioinformatics alignment method to determine personalized proteomic profiles for membrane proteins in human tissues. This strategy provided accurate (6% error) and reproducible (34% relative S.D.) quantification of three independently purified membrane fractions from the same human colorectal cancer (CRC) tissue. Using CRC as a model, we constructed the personalized membrane protein atlas of paired tumor and adjacent normal tissues from 28 patients with different stages of CRC. Without fractionation, this strategy confidently quantified 856 proteins (≥2 unique peptides) across different patients, including the first and robust detection (Mascot score: 22,074) of the well-documented CRC marker, carcinoembryonic antigen 5 by a discovery-type proteomics approach. Further validation of a panel of proteins, annexin A4, neutrophils defensin A1, and claudin 3, confirmed differential expression levels and high occurrences (48-70%) in 60 CRC patients. The most significant discovery is the overexpression of stomatin-like 2 (STOML2) for early diagnostic and prognostic potential. Increased expression of STOML2 was associated with decreased CRC-related survival; the mean survival period was 34.77 ± 2.03 months in patients with high STOML2 expression, whereas 53.67 ± 3.46 months was obtained for patients with low STOML2 expression. Further analysis by ELISA verified that plasma concentrations of STOML2 in early-stage CRC patients were elevated as compared with those of healthy individuals (p < 0.001), suggesting that STOML2 may be a noninvasive serological biomarker for early CRC diagnosis. The overall sensitivity of STOML2 for CRC detection was 71%, which increased to 87% when combined with CEA measurements. This study demonstrated a sensitive, label-free strategy for differential analysis of tissue membrane proteome, which may provide a roadmap for the subsequent identification of molecular target candidates of multiple cancer types.