Clinical Proteomics

Background: Pre-operative triage of adnexal masses with serum CA-125, HE4, and ultrasound (O-RADS) has limited accuracy, contributing to unnecessary surgery. We develop and validate a first of its kind, biomarker panel-free plasma proteomic classifier that distinguishes malignant from benign adnexal masses from a single blood draw, operating directly on discovery-mode mass spectrometry. Methods: In this multicenter prospective observational study, plasma from women with adnexal masses and from healthy controls at a large urban academic center (discovery and four validation cohorts) and an external academic center from another Urban-suburban area (external validation) was analyzed by data-independent-acquisition mass spectrometry (Orbitrap Astral), quantifying >1,000 proteins per sample. Cancerous (malignant, borderline adnexal mass, or metastasis to ovary) versus benign adnexal masses were classified with ADAPT-MS (Adaptive Diagnostic Architecture for Personalized Testing by Mass Spectrometry), which retrains an ElasticNet model for each sample based on the proteins measured. The primary outcome was discrimination (area under the receiver operating characteristic curve [AUC]). Pre-operative O-RADS scores and serum CA-125 were compared and combined with proteomics on matched subsets. Findings: Using a discovery cohort (n=1062), an ADAPT-MS classifier, validated across five independent cohorts (n=623), achieved a summary AUC of 0.835 (cohort range 0.779-0.904). Among patients with pre-operative O-RADS (n=231), proteomics (AUC 0.852) was comparable to O-RADS alone (0.863) and combining them raised discrimination to 0.922. Proteomics outperformed serum CA-125 (0.915 versus 0.777; n=125) and proteome-derived ROMA-like and two-marker surrogates. Biologically, the classification-relevant proteins are enriched for host response proteins rather than tumor cell associated protein changes, comprising ECM remodeling, complement and innate immune response pathways. This explains good detection of tumor with heterogenous pathology and also metastasis to ovary. Interpretation: A single plasma proteomic measurement interpreted using adaptive machine learning discriminates malignant from benign adnexal masses with good internal and external validation, and is complementary to ultrasound, warranting prospective evaluation as a panel-free diagnostic adjunct to imaging.

Proteolysis can be completed in a single-step, rapid (5-30 minutes) reaction using hyperthermoacidic archaeal (HTA) proteases. Protocol simplicity and stability of HTA proteases in ambient storage conditions removes many logistical barriers to widespread deployment of proteomic analysis. HTA protease cleavage selectivity is complicated but quantitatively reproducible and generates peptides that negligibly overlap with trypsin. Tryptic and HTA protease peptides emphasize different proteins and provide complementary coverage of amino acid sequence and post-translational modification (PTM) sites. Digestion of a sample's separate aliquot with HTA protease is a minor addition to trypsin-based workflows, but LC-MS analysis of both sets of peptides is not practical for large cohorts. This manuscript evaluates the identifications and reproducibility of LC-MS analysis of tryptic and HTA protease peptide mixtures. Adding HTA protease peptides to 20% increased total protein identifications by 24% on average in neat plasma and identified 216 additional non-artifact modifications compared to conventional trypsin-only bottom-up proteomic analysis. This approach is reproducible and does not increase sample preparation or analysis time. Peptides from each protease present independent opportunities to quantify overlapping proteins and a quantification model was developed in which the protease with the most complete data is selected for each protein while the second protease is used to infer missing values. This cross-protease inference is more accurate than the alternative minimum, mean, or nearest neighbor imputations. Finally, biomarkers discovered through HTA protease peptides are amenable to rapid HTA protease only proteomic workflows thus streamlining translation from discovery to fieldable applications.

In bottom-up proteomics, protease specificity determines peptide length and LC/MS/MS detectability; in N-terminomics, this is critical because each protein N-terminal peptide is unique and cannot be substituted by other peptides. CHromatographic AMplification of Protein N-terminal peptides (CHAMP-N) combines LysargiNase digestion with StageTip-based strong cation exchange chromatography (SCX), removing internal peptides by exploiting their newly generated N-terminal Lys/Arg residues and thereby selectively isolating protein N-terminal peptides. However, LysargiNase cleavage near the protein N-terminus can generate excessively short peptides. Here, we developed two complementary CHAMP-N approaches that modurate N-terminal peptide length while retaining the charge-based SCX separation principle. LysN digestion avoided cleavage at N-terminal-proximal Arg residues, whereas D3-acetylation blocked Lys cleavage by LysargiNase to produce ArgN-like specificity. Both approaches extended otherwise short N-terminal peptides while preserving N-terminal basic residues on internal peptides. Conventional LysargiNase, LysN, and ArgN-like CHAMP-N generated complementary sets of N-terminal peptides. Integration of the three workflows identified 2,552 protein N-termini, approximately twice the number identified by conventional CHAMP-N alone. The median relative standard deviations of N-terminal peptide intensities were approximately 15% for all three workflows, indicating comparable technical reproducibility. These results demonstrate that coordinating protease cleavage specificity with SCX selectivity reduces sequence-dependent losses and expands N-terminome coverage while preserving the simplicity and reproducibility of CHAMP-N.

Kidney disease affects over 850 million people and leads to degradation of the kidney cortical compartments and kidney failure. Interstitial fibrosis and tubular atrophy (IFTA) is a universal feature of chronic kidney disease, yet the sequence of tubulointerstitial matrix protein (matrisome) changes that underlie the development and progression of IFTA are not fully understood. In particular, post-translational modifications which may stabilize key matrisome components and result in irreversible and functionally detrimental interstitial fibrosis or tubular basement membrane thickening, have yet to be characterized. Post-translational modifications (e.g., acetylation) have been shown to contribute to kidney diseases, yet they have not been investigated in chronic kidney disease progression. Herein, we established feasibility of matrisome protein quantification in the tubulointerstitium, with concurrent matrisome protein post-translational modification quantification, in archived diagnostic human kidney biopsies, using laser capture microdissection, liquid chromatography tandem mass spectrometry (LC-MS/MS), and bioinformatics analysis. In this pilot study, we focused on two conditions associated with IFTA progression: diabetic nephropathy and kidney transplantation. We examined both healthy and chronically damaged tubulointerstitium in two human kidney biopsies, one representing each condition. We identified signature protein markers of tubulointerstitium (e.g., UMOD, SLC12A1) in abundance, confirming sample composition, and observed uniformity in protein abundance across replicates from the same biopsies. We found increases in certain collagen isoforms (COL1A1, COL6A3) and other ancillary proteins in areas of IFTA. We also successfully quantified post-translational modifications, confirming that hydroxylproline was the most abundant modification in healthy tubulointerstitium from the kidney transplant biopsy, and that hydroxyllysine was also highly abundant and could be confidently quantitated on two collagen isoforms (COL1A1, COL4A2). Lastly, we performed power analysis to determine the required sample size to achieve significant power and false discovery rate. Several proteins exhibited large Hedge effect sizes in the tubulointerstitium (ANPEP, PRODH2), indicating their potential usefulness as fibrosis signatures. Our study is a first step toward comprehensive characterization of tubulointerstitial matrisome changes, including post-translational modifications seen with IFTA progression. Subsequent studies will potentially reveal new molecular markers of key early events in chronic kidney disease progression, and new therapeutic targets for mitigating this process. Data are available via ProteomeXchange identifier PXD079025.

Detecting low-abundance bacterial peptides in samples dominated by human proteins using liquid chromatography tandem mass spectrometry (LC MS/MS) is challenging. Bacterial peptides are typically far less abundant than host-derived peptides, and sequence similarity between bacterial and human peptides can complicate confident attribution to a bacterial origin. To address this problem, we developed and benchmarked an in silico spectral library protein identification workflow for detecting Staphylococcus aureus USA300 peptides against a background of A375 cells. Peptide and protein identifications obtained at a 1% peptide-level false discovery rate were compared between the in silico spectral library search and a conventional sequence database search across a defined bacterial dilution series containing 0, 1%, 2%, 10%, 25%, 50%, and 100% bacterial protein. A human sequence similarity filter was applied to identifications from both search strategies to reduce the likelihood of retaining human derived or human like peptide sequences. Both search strategies showed concentration dependent bacterial detection and shared a practical detection limit of approximately 1 to 2% bacterial protein but exhibited distinct performance profiles. The in silico spectral library search produced fewer bacterial-like background identifications in the human-only control, whereas MS2Rescore substantially increased spectral library identifications at higher bacterial concentrations, exceeding database search at 100% bacterial protein. The partially overlapping peptide sets identified by the two strategies indicate that they provide complementary coverage of the bacterial proteome.

Blood-based metabolomic profiling has been widely investigated for breast cancer (BC) detection; however, clinical implementation remains limited due to variability in sample handling, analytical reproducibility, and overfitting during statistical analysis. We established a plasma GC/MS metabolomics workflow for discriminating BC from healthy controls (HC) using conventional machine-learning algorithms. Plasma samples (n = 360; BC = 180, HC = 180) were collected prospectively under standardized preanalytical conditions before surgery and the initiation of systematic anticancer therapy and analyzed using a quantitative GC/MS platform with automated derivatization. Feature selection and model development were conducted using three machine-learning (ML) algorithms (Lasso logistic regression (LR), random forest classifier (RFC), and support vector machine (SVM)). A total of 45 metabolite candidate biomarkers were identified, and the optimal number of metabolite features for each algorithm was estimated by a recursive feature elimination (RFE)-based strategy. The best-performing models achieved area under the ROC curve values (AUC) of 0.910 (LR), 0.893 (RFC), and 0.843 (SVM). We selected prioritizing candidate biomarkers consistently expressed across the multi-algorithm pipeline. A bagging ensemble model improved stability (AUC = 0.911) and reduced false-positive predictions in the independent HC dataset. In addition, model stability with respect to false-positive predictions was assessed using an independent HC cohort (n = 15) that was collected at a separate institution. These results indicate that a plasma metabolomics workflow combined with conventional multi-algorithm ML, algorithm-specific feature selection, and independent assessment provides stable discrimination between BC and HC in a moderately sized cohort.

High throughput sequencing necessitates robust computational tools to contextualize complex transcriptomic data. In this study, we analyzed RNA Seq data from peripheral blood of bacterial pneumonia patients and healthy controls (GEO GSE161731) to investigate systemic molecular responses to infection. Differential gene expression analysis identified significantly upregulated genes in pneumonia. We used GANGO for Gene Ontology (GO) term enrichment and BioFunctional for hierarchical exploration, providing an initial overview of affected biological processes. To deepen interpretation, we leveraged CURIE (www.curieai.tech ), an AI plat-form that cross-references GO terms with scientific literature. CURIE generates narrative explanations, link-ing molecular mechanisms, experimental evidence, and publications to each biological process, moving be-yond simple enrichment. Our combined analysis revealed a coordinated upregulation of pathways involved in DNA repair and replication, transcriptional activation, RNA processing, translation, ribosome biogenesis, mitochondrial respiration, and cofactor binding. By integrating these findings within a broad, literature-supported framework, we highlighted how multiple cellular compartments and processes collaborate to mount an effective response to infection-induced stress. This study underscores the potential of combining classical statistical enrichment with AI based interpretation for a holistic understanding of host-pathogen in-teractions at the transcriptomic level.

Abstract Background: High-grade serous ovarian cancer (HGSC) is characterized by extensive genomic heterogeneity and frequent alterations that promote replication stress (RS). Although RS has emerged as a promising therapeutic vulnerability, it remains unclear whether distinct RS-associated genomic alterations generate similar tumour states or drive unique biological programs and therapeutic dependencies. Methods: We developed and characterized a panel of syngeneic HGSC models representing recurrent RS-associated alterations, including overexpression of Ccne1, Brd4, Myc, Ndrg1, Pik3ca, and Brca1 heterozygosity, alongside previously established Pten/Nf1-deficient and Myc overexpressing models. Tumour fitness, transcriptomic programs, histopathology, extracellular matrix composition, immune organization, and therapeutic vulnerabilities were assessed using subcutaneous (s.c.) and intraperitoneal (i.p.) allografts, RNA sequencing (RNAseq), mass spectrometry (MS), cyclic-immunofluorescence (Cyc-IF), and preclinical therapeutic studies. Results: RS-associated genotypes exhibited marked differences in tumour fitness and generated distinct tumour ecosystem states despite sharing a common genetic background. Tumours harbouring BRCA1 heterozygosity displayed reduced tumour fitness, mesenchymal features, extracellular matrix remodelling, stromal expansion, and immune-excluded phenotypes. In contrast, tumours co-expressing Ccne1 and Brd4 exhibited the highest RS level together with strong activation of DNA damage response pathways and limited immune infiltration. Immune characterization identified immune-desert, immune-suppressed, and immune-excluded states across models, demonstrating substantial genotype-dependent immune heterogeneity. Elevated CD47 expression emerged as a defining feature of the Ccne1/Brd4 tumour state and was confirmed in human HGSC specimens with high CCNE1 and BRD4 protein expression. Therapeutic targeting of ATR and CD47 in intraperitoneal Ccne1/Brd4 allografts reduced tumour burden and induced histopathological responses compared with either monotherapy alone, supporting the translational relevance of this RS-high tumour state. Conclusions: Distinct RS-associated alterations do not converge on a common HGSC phenotype but instead generate diverse tumour ecosystem states with unique effects on tumour fitness, stromal remodelling, immune organization, and therapeutic vulnerabilities. These findings highlight the importance of considering the genetic origin of RS when developing therapeutic strategies and establish a framework for investigating genotype-specific vulnerabilities in HGSC.

Mass spectrometry based immunopeptidomics is limited by peptide identification. HLA class I ligands are non-tryptic, frequently carry low charge states and often fragment incompletely, yet nearly all workflows rely on a single activation method, higher-energy collisional dissociation (HCD). Unlike proteomics, where a protein is inferred from several peptides, in immunopeptidomics the peptide is itself the analyte and the candidate therapeutic target, so a ligand that is not identified is a target that is not found. Here we evaluate whether two alternative activation chemistries, ion trap collision-induced dissociation (CID) and infrared multiphoton dissociation (IRMPD), recover ligands that HCD does not. To our knowledge this is the first application of either method to HLA class I immunopeptidomics, made practical by the integrated infrared laser of the Orbitrap Tribrid Apex MultiOmics. Four HLA-typed cell lines were analysed in technical triplicate against the canonical human proteome and a Ribo-seq-derived cryptic ORF database. Because fragment-intensity predictors are trained almost exclusively on HCD spectra, AlphaPeptDeep was fine-tuned separately for each method; fine-tuning and Percolator rescoring increased unique peptide identifications in every cell line and method combination (median +17.1% canonical, +54.5% cryptic), and fine-tuned models transferred across cell lines with a median loss of 0.4%, placing the three methods on a common analytical footing. On that footing the three methods yielded immunopeptidomes of similar size and composition, but each recovered peptides the others missed. HCD produced the most complete fragment-ion ladders and recovered more peptides from the basic-anchor alleles A*11:01 and A*33:03, both members of the A3 supertype. CID and IRMPD recovered a larger share of alleles without a basic C-terminal anchor and, for peptides containing no lysine, arginine or histidine, produced higher search scores and more matched b/y ion current in three of four cell lines. Each method therefore contributed exclusive peptides, and at a fixed budget of three LC-MS/MS runs, one run per method yielded on average more unique peptides than triplicate runs of a single method in all four cell lines (+0.4% to +10.9%). Of 239 novel canonical predicted HLA binders in the two cancer cell lines (HCT116 and MDA-MB-231) absent from the Immune Epitope Database, 55% were detected by only one method; nine derive from cancer-associated antigens. CID and IRMPD can therefore provide sequences that may not be identified by HCD alone, or with better spectral quality for a chemically defined subset of ligands. Their value lies alongside HCD rather than in its place, and realising it will require fragment-intensity models and acquisition schemes built for these chemistries rather than adapted from HCD.

Background: Tuberculosis (TB) remains the leading infectious disease cause of mortality worldwide. A subset of individuals exposed to Mycobacterium tuberculosis (Mtb) remain TST/IGRA-negative despite sustained exposure, termed "TB resisters". The biological mechanisms underlying this resistance remain poorly understood. We applied untargeted high-resolution metabolomics to determine whether systemic metabolic profiles differ between TB resisters and matched Mtb-infected close contacts. Methods: We compared untargeted high-resolution plasma metabolomics using liquid chromatography mass spectrometry between 269 TB resisters and 269 matched Mtb-infected close contacts, enrolled across Brazil, India, and South Africa. TB resisters were defined as highly exposed close contacts who slept in the same room or spent at least 5 hours indoors per day with an infectious pulmonary TB index patient, but remained TST/IGRA-negative. Mtb-infected close contacts were those who were TST/IGRA positive. Metabolome-wide association studies (MWAS) were conducted using paired t-tests. Results: There were 1,787 features with nominal differences between TB resisters and matched Mtb-infected contacts (p < 0.05). Pathway enrichment identified fructose and mannose metabolism and bile acid biosynthesis in the overall cohort, while lipid-related pathways were enriched in Brazil. Among metabolites with confirmed chemical identities, glyceric acid concentrations were higher in TB resisters by 0.53 uM, whereas hydroxyproline was lower by 0.90 uM, butyrobetaine was lower by 0.06 uM, and homogentisate was lower by 0.0036 uM at false discovery rate of 0.2. Conclusions: Plasma metabolomic profiles differed between TB resisters and matched Mtb-infected close contacts. These findings indicate that systemic metabolic differences are associated with resistance to Mtb infection.

The rapid expansion of mRNA-based medicines has driven demand for robust, high-resolution analytical methods capable of characterising critical quality attributes including sequence identity, 5[prime] capping efficiency, and poly(A) tail length and heterogeneity. Here, we present two complementary liquid chromatography-mass spectrometry based workflows for mRNA characterisation, both based on site-specific digestion with the endoribonuclease MazF from Escherichia coli. A middle-up approach, optimised via specific 5[prime]-ACA cleavage enables rapid characterisation via oligoribonucleotide mass mapping providing complete sequence coverage and simultaneous assessment of 5[prime]-capping efficiency and 3[prime]-poly(A) tail length and heterogeneity. This workflow was successfully applied to three different mRNA constructs, NLuc, eGFP and FLuc mRNAs, achieving sequence coverages of 100%, 87%, and 74%, respectively. A complementary bottom-up approach, employing less specific 5[prime]-ACX cleavage enables tandem mass spectrometry-based sequencing of shorter oligoribonucleotides and detailed characterisation via mRNA sequence mapping. Applied to SARS-CoV-2 Spike Protein mRNA, this method yielded 50% sequence coverage based on unique oligoribonucleotides with MazF alone, extended to 90% by combining MazF, partial RNase T1, and partial RNase U2 digests. Together, these workflows provide a flexible, orthogonal platform for both high-throughput quality assessment and in-depth primary sequence characterisation of mRNA vaccines and therapeutics.

Introduction: Neoepitopes form the basis of tumor-specific immune responses. Tissue biopsy, the primary source for neoepitope detection, is limited and invasive. Therefore, we aimed to identify neoepitopes by mutanome-guided immunopeptidomics from plasma of cancer patients. Methods: Mass spectrometry (MS) data analysis of HLA ligands from plasma (n = 4) was guided by patient-specific mutanomes of cell-free DNA (cfDNA) from plasma or tumor genomic DNA (tgDNA) from tissue. Matched tumor tissue and healthy donor plasma served as controls. Neoepitopes were validated with synthetic peptides, and immunogenicity was assessed using IFN-gamma ELISpot and intracellular cytokine staining. Results: Wild-type immunopeptidomes from tissue and plasma overlapped by 58%, with 91% of plasma HLA ligands rediscovered in tissue. 13 out of 15 tumor-associated HLA ligands detected in plasma were rediscovered in the matching tissue. However, no neoepitopes in plasma were identified by immunopeptidomics guided by cfDNA mutanomes, likely reflecting the limited overlap between cfDNA and tgDNA mutanomes (15%). Using the tgDNA mutanome as a complementary reference, two neoepitopes were detected in one patient's plasma, albeit at the MS detection limit. Both neoepitopes were also discovered in tissue, along with three tissue-exclusive neoepitopes. Two tissue-exclusive neoepitopes induced antigen-specific T cell responses in healthy donor PBMCs. Conclusion: In summary, plasma immunopeptidomics enables profiling of HLA ligands from wild-type proteins, including TAAs. In principle, neoepitope detection from plasma at the peptide level is feasible, but tissue remains the gold standard for variant calling and neoepitope identification. Improved detection methods may enable minimally invasive approaches in the future.

Pediatric endoscopy is performed only when clinically indicated, limiting access to healthy duodenal tissue. Biopsies with duodenal no pathologic abnormality (NPA) are often used as controls despite the presence of symptoms or inflammatory disease found elsewhere in the gastrointestinal (GI) tract. We characterized pediatric duodenal NPA tissue across clinical, histologic, cellular, and transcriptomic domains. Methods Archival duodenal NPA biopsies were obtained with clinical metadata and hematoxylin-and-eosin whole-slide images (WSIs). Duodenal mRNA-seq data were analyzed from a subset of patients with duodenal NPA. Clinical metadata and WSIs underwent machine-learning analysis, cell populations were quantified from WSIs, and RNA-seq data underwent differential expression and pathway-enrichment analyses. Results The primary cohort included 195 patients with duodenal NPA. Comparisons between patients with non-duodenal GI disease and those with no GI disease showed differences in inflammatory biomarkers and follow-up utilization. Unsupervised clinical clustering identified three clusters with partial enrichment for IBD with colonic inflammation and Eosinophilic Esophagitis (EoE) with esophageal inflammation. Supervised clinical classification showed modest discrimination. WSI clustering showed limited disease-status discrimination, and cell quantification showed no significant group differences. In the separate RNA-seq cohort of 43 patients, differential-expression and pathway-enrichment analyses identified transcriptional and pathway-level differences between disease-status groups. Conclusions This multi-level characterization indicates that pediatric duodenal NPA tissue should not be treated as a uniform control category. Clinical metadata and transcriptomics revealed clinical and molecular heterogeneity, while histologic and cell analyses showed limited disease-status separation, supporting a refined definition of control tissue.

Ubiquitination is a post-translational modification with well described roles in the regulation of immune cell function, such as dendritic cells (DCs). DCs are crucial initiators and regulators of T cell-mediated immune responses. In this study, we use activity-based probes and proteomics to systematically profile DUBs in a DC line. Using the lysate-based biotin(bio)-Ubiquitin(Ub)-propargylamide(Prg) probe and proteomics, the activity and abundance of deubiquitinases was measured in unstimulated and CpG-stimulated MutuDCs to simulate responses to bacteria. This analysis was complemented with profiling using the cell-permeable DUB activity-based probe IMP-2373. The combined analysis with bio-Ub-Prg and IMP-2373 provided insights into the dynamics of DUB activity and abundance in DCs following stimulation and identified several DUBs for further study.

Central carbon metabolism coordinates energy production and carbon skeleton allocation across photosynthetic and heterotrophic organisms. Here, we established an optimized targeted metabolomics workflow for Chlamydomonas reinhardtii and Arabidopsis thaliana using a single-run HILIC-MS/MS assay with alternating positive and negative ionization. Our method provides robust multi-class coverage across 21 amino acids, 11 carboxylic acids, 4 sugars, 3 adenine nucleotides, 2 nucleotide sugars, 18 phosphorylated compounds in CCM, and Acetyl-CoA in a 20-minute run using both ESI modes. An optimal concentration of medronic acid with a BEH Amide column at high pH prevented metal adsorption, and low quantification limits (10 to 50 nM) for labile intermediates such as PEP, ATP and RUBP were achieved. Retention time variations remained below 1.3%, and over 90% of analytes showed high linearity (R2 > 0.990) across dynamic ranges spanning two to three orders of magnitude (typically 0.01-0.5 uM up to 15-20 uM). For sample preparation, glass-fiber filters (GF/C) outperformed nitrocellulose for harvesting C. reinhardtii. For A. thaliana, it was proved that no further grinding with micro pestle and extended cooling were necessary. A two-phase chloroform/methanol/water extraction was essential, as a single-phase extraction co-extracted green chlorophyll, which heavily suppressed phosphorylated signals. In Arabidopsis, glutamate, fumaric acid and citric acid were the dominant primary pools, indicating the importance of dilution and column washing for preventing carryover, and sensitive intermediates were reliably quantified at physiological levels. Together, this unified analytical and extraction strategy enables robust, artifact-free, high-throughput profiling of central metabolism in photosynthetic organisms.

Prostate cancer (PC) bone metastases (PCBM) are a debilitating disease morbidity that primarily affects the axial skeleton. PCBM are predominantly osteosclerotic with mixed osteosclerotic/osteolytic features, including irregular bone deposition, loss of collagen alignment and increased porosity that mimic features of woven bone formed during development and repair. We hypothesize that progression of PCBM is driven by interactions between tumor cells and the bone microenvironment and is associated with changes in the stromal compartment. We performed tandem mass tag mass spectrometry proteomics on lumbar vertebral specimens from 32 cadaveric PCBM specimens and 3 age-matched controls. Unsupervised clustering segregated specimens into three clusters associated with benign, mixed, and osteosclerotic lesion states, characterized by distinct immune-, tumor-, and osteogenic-associated protein signatures. Histological analysis showed tumor infiltration was accompanied by marrow fibrosis and increased collagen III deposition. We identified a marker signature associated with osteosclerotic remodeling and mapped this signature to independent metastatic PC transcriptomic datasets. The proteomics-derived markers reproduced similar molecular groupings at RNA level, and robustly segregated bone PCBMs from soft tissue metastases PC in RNA data. Together, these findings define molecularly distinct states of tumor-driven changes in PCBM that extend beyond histological classification and may inform biomarker development and therapeutic stratification.

Key plasma protein, including albumin, complement system proteins, and coagulation factors circulate in biochemically and functionally distinct isoforms, or proteoforms. Despite the importance of these proteoforms being widely accepted, they are challenging to resolve systematically, and as a consequence, we still lack a more complete picture of the total proteoform diversity of the human plasma proteome. Here, we generated 3,114 intact protein fractions by extensively separating plasma according to isoelectric point and molecular weight. We then analysed these fractions by liquid chromatography data-independent acquisition mass spectrometry (DIA-MS) upon tryptic digestions, and combined the resultant data into a multidimensional, proteoform resolved atlas of the human plasma proteome. We report that 5,077 unequivocally detected plasma proteins circulate at least 179,049 distinct proteoforms, of which we detect 44,083, with a median of 11 proteoforms per protein, at high confidence. Substantially increasing the number of experimentally confirmed proteoforms, our atlas implies that the canonical protein forms account for less than half of the total protein mass in plasma. We detect the greatest proteoform diversity among the most abundant plasma proteins and those functioning in nutrient transport, complement and coagulation. Finally, our data clarifies discrepancies between the Olink and SomaScan proteomic platforms, as according to our atlas, these techniques provide correlated values for relatively abundant proteins with low to medium proteoform diversity, but struggle in the quantification of low-abundant proteins and on those which exist in many proteoforms. In sum, our data reveals extensive proteoform diversity of the human plasma proteome, and implies that proteoform diversity represents an enormous untapped reservoir for biomedical research and biomarker discovery.

Combinatorial therapies are widely used in the treatment of acute myeloid leukemia (AML) to address disease heterogeneity, adaptive resistance, and rewired signaling and metabolic states. Yet drug prioritization remains largely guided by clinical or phenotypic evidence, while the molecular mechanisms underlying effective drug combinations remain incompletely defined. To narrow this gap, we developed Combinatorial high-ratio Partial proteolysis with reference PRoteome Analysis (CoPPRA), a structural proteomics workflow based on limited proteolysis of cell lysates that profiles drug-associated changes in regional protein accessibility at peptide-level resolution. Here, we applied CoPPRA to ruxolitinib and ulixertinib, individually and in combination, in AML-related cell lysates. Our findings extend conjunctive targeting (CT), a recently proposed mechanism of combinatorial drug action in which combined exposure produces protein targeting patterns not observed with either drug alone. Previously identified through combination-associated changes in protein solubility/stability, CT is examined here at peptide-level resolution through regional differences in proteolytic accessibility. The ruxolitinib-ulixertinib combination produced broad peptide-level accessibility changes, including a subset meeting the predefined criteria for CT. CT candidates predominantly exhibited regional accessibility changes, with altered peptide regions occurring against comparatively small changes across the remaining quantified peptides from the same proteins. MAP2K1 and ATP6V1G1 showed pronounced differences between overlapping peptide sequences, highlighting localized variation in combination-associated accessibility, including an ATP6V1G1 peptide mapping to an annotated helical region. Combination-associated increases in peptide signals were also observed in PIK3R1, BRD4, and PTPN11, linking regional accessibility changes to signaling and transcriptional regulators relevant to AML. Functional enrichment and network analyses further implicated nucleotide and glucose metabolism, ficolin-1-rich granules, ribosome-associated processes, and phagocytic vesicles. These results extend conjunctive targeting from protein-level solubility/stability changes to regional differences in proteolytic accessibility, showing that combination-associated effects can be concentrated within specific peptide regions rather than distributed uniformly across proteins. More broadly, CoPPRA provides a peptide-resolved approach for investigating the molecular features of combinatorial drug action and prioritizing protein regions for subsequent mechanistic validation.

Pancreatic ductal adenocarcinoma (PDAC) arises within an extensively remodeled microenvironment, yet tissue markers that distinguish invasive carcinoma from precursor lesions remain limited. Integrating four single cell RNA sequencing atlases comprising 324,662 cells from 26 patients, we identified cancer associated fibroblast expansion, accumulation of C1QA+ immunoregulatory macrophages, and depletion of a dendritic cell compartment expressing plasmacytoid markers. Fibroblasts were enriched for extracellular matrix and PI3K AKT programs and macrophages for an antigen-presentation program. We then examined the protein level correlates of these populations in a tissue proteomic reference framework of 356 samples spanning non diseased pancreas, normal-adjacent tissue, low and high grade intraductal papillary mucinous neoplasms (IPMNs), IPMN associated invasive carcinoma, and PDAC, prioritizing proteins that changed monotonically across the six histopathologic reference groups or marked tumor enriched cell types. Despite depletion of plasmacytoid dendritic cells, the principal source of type I interferon, the interferon stimulated protein ISG15, expressed by both fibroblasts and macrophages, increased monotonically toward PDAC. In exploratory cross-validated benchmarking of PDAC versus low and high grade IPMN, excluding IPMN-associated invasive carcinoma (145 versus 56 samples), CATB+EPIPL achieved an AUC of 0.98, with a sensitivity of 92% at an achieved specificity of 92%, exceeding protein abundance models of diagnostic IHC-associated markers (S100P+IMP3+maspin+MSLN, 0.79) and refitted published IPMN progression combinations (<0.91). Against non diseased pancreas, candidate and comparator panels performed similarly (AUC 0.99 1.00). These biologically informed tissue protein panels warrant independent validation, including immunohistochemical evaluation of EUS guided fine needle biopsy specimens and measurement in blood.

Cerebrospinal fluid (CSF) contains brain derived proteins that can be leveraged as biomarkers for Alzheimer's disease (AD) and other neurodegenerative diseases. However, converting protein intensities to a disease signal remains a challenge because physiology driven changes to the localization of proteins across CSF subcompartments are obscured when CSF is analyzed alone. Where subcompartment data, such as that from extracellular vesicles (EVs), are also measured, separate but parallel analyses of EVs and CSF can obscure their shared disease signal. Here, we present an integrated matrix factorization approach that combines proteomic signals from matched CSF and phosphatidylserine exposed EVs and nanoparticles (EVps). In a ROS/MAP cohort of 48 participants, integrating CSF and EVps signals improved biological concordance across compartments and increased the number of candidate proteins distinguishing participants with AD from controls. Notably, the joint signal identified as many brain related biomarker candidates as studies using substantially larger cohorts. By demonstrating that integration of CSF and EVps proteomes maximizes information from limited samples, this framework offers a path toward more efficient biomarker discovery across two compartments increasingly leveraged in neurodegenerative disease research.