Intracellular protein editing enables the incorporation of non-canonical amino acids into endogenous proteins

      Proximity proteomics enables integrated profiling of the single-cell-type spatial proteome in complex tissues

      Sex differences in the genetic regulation of the human plasma proteome

      Integrated transcriptomic, proteomic and metabolomic analysis reveals inflammatory signatures and pathways shared across severe falciparum malaria subtypes

      Cerebrospinal fluid proteomics reveals the molecular signature of frontotemporal lobar degeneration

      Pre-analytical bias factors in magnetic bead-enriched plasma proteomics

      Non-invasive prognostic classification of intratumoural heterogeneity in liver cancer, with multi-omics characterisation and therapeutic implications

      Scaled profiling of the endogenous single-cell proteome in the rat hippocampus

      Pan-cancer proteomic atlas: a mass spectrometry-based panorama of tumour biology, biomarker and therapeutic target discovery

On 1 May 2025, the teams of George M. Burslem and Ophir Shalem at the University of Pennsylvania delivered a breakthrough advance — they developed a revolutionary "protein editing" technology that, for the first time, achieves the precise installation of non-standard amino acids and chemical labels into endogenous proteins in living mammalian cells. At the heart of the technology lies the precise regulatory system of split inteins. The team selected two pairs of orthogonal split inteins, Gp41-1 (with a splicing half-life of about 5 seconds) and AvaN:Npuc (with a splicing half-life of about 1 minute), and constructed a tandem trans-splicing system. Using CRISPR-Cas9, the intein acceptor was inserted precisely at a specified site of the target protein; it contains the two split domains Gp41-1N and NpucC, with the mNeonGreen211 fluorescent protein and a Myc tag linked in between as temporary markers. The donor protein is expressed recombinantly in Escherichia coli and carries the target insertion sequence (such as an HA tag or a non-natural amino acid) together with the corresponding Gp41-1C and AvaN domains. Once the donor protein is delivered into the cell by electroporation or lipid nanoparticles, the two sets of split inteins complete trans-splicing in succession, achieving precise embedding of the functional module while completely removing the intein structure, thereby ensuring the structural integrity and functional realisation of the target protein.

▼

Spatial proteomics can carry out in-depth analysis of tissue structure through laser microdissection coupled with mass spectrometry (LMD-MS) and antibody-based imaging. However, the trade-off between sampling precision, detection throughput and proteome coverage still limits the widespread application of existing strategies. In response, the team of Professor Tian Ruijun at the Southern University of Science and Technology has proposed an innovative proximity labelling technology for spatial proteomics (PSPro). The method combines highly specific antibody-targeted biotinylation with an efficient affinity purification workflow to achieve precise capture of the proteome of specific cell types in a single tissue section at submicron resolution. Through parameter optimisation and systematic validation, PSPro demonstrated excellent performance in comparison with conventional flow cytometry (FACS) and laser microdissection proteomics workflows. In further application studies, the team applied PSPro to tumour and spleen tissue sections and successfully enriched thousands of proteins from 10 different cell types, covering a range of known cell markers. By integrating LMD technology into the PSPro workflow, the researchers were able to compare the spatial proteomes of cell subpopulations within the same tissue section, revealing the heterogeneity of protein expression between tumour cells and immune cells in pancreatic cancer tissue. Overall, with a simple and scalable workflow, PSPro extends the traditional "antibody–epitope" recognition paradigm into a new "antibody–cell-type proteome" model suited to spatial biology research, providing powerful technical support for proteomic studies of complex tissue microenvironments.

Sex differences in the genetic regulation of the human plasma proteome

Many diseases show marked sex differences in onset, progression and prognosis, yet the biological mechanisms behind them remain unclear. A research team from Queen Mary University of London systematically analysed sex differences in the genetic regulation of the plasma proteome (covering more than 5,800 proteins) across two large population cohorts, including 30,307 women and 26,058 men. The results showed that more than two-thirds of proteins differed significantly in expression level between the sexes. By comparison, however, the genetic effects on protein levels were highly conserved between men and women: only 103 sex-differentiated protein quantitative trait loci (sd-pQTLs) were identified, accounting for 2.9% and 0.3% respectively of the loci detected by the antibody platforms used. This indicates that the influence of sex on the genetic regulation of proteins is on the whole rather limited. Among these sd-pQTLs, around one-third showed sex-specific regulatory patterns: 30 loci exerted an effect in only one sex, while the CDH15 locus showed regulatory trends in opposite directions in the two sexes. Further phenome-wide association analysis incorporating 365 phenotypic measures from UK Biobank found no clear evidence that these sd-pQTLs explain sex-related differences in disease risk. The study shows that although protein expression levels differ widely between the sexes, their genetic regulation is broadly consistent across sexes. This finding provides an important reference for genetics-based drug target discovery and validation.

The mechanism of clinical progression in Plasmodium falciparum infection has long been one of the central questions in malaria research. A joint research team from the University of Science, Techniques and Technologies of Bamako in Mali and the University of Maryland School of Medicine carried out a systematic multi-omics integration analysis on 79 children from Mali, covering transcriptomic, proteomic and metabolomic data. They made an in-depth comparison of the molecular profiles of patients with cerebral malaria (CM), severe malarial anaemia (SMA) and CM combined with SMA against those of non-severe malaria controls, revealing the shared and subtype-specific inflammatory mechanisms across different severe malaria subtypes. The study found that MMP8, IL1R2 and ARG1 were consistently and significantly up-regulated in all severe malaria subtypes, indicating the presence of a common systemic inflammatory signature. Among them, tissue inhibitor of metalloproteinase 1 (TIMP1) showed the most marked expression in CM; together with the high expression of MMP8/MMP9 this revealed the key role of the metalloproteinase pathway in the pathological process within the central nervous system. In addition, levels of L-arginine metabolites were significantly reduced in children with cerebral malaria; combined with the increase in ARG1, this suggests a possible mechanism whereby disturbance of arginine metabolism leads to restricted cerebrovascular relaxation. Through cross-validation across multiple omics layers, this study systematically characterised the inflammatory cascade in severe malaria, deepening our understanding of disease mechanisms and opening up new directions for targeted therapeutic strategies and the development of inflammation-related biomarkers.

The pathophysiological mechanisms of frontotemporal lobar degeneration (FTLD) and the corresponding biomarkers still await deeper exploration. The team of Rowan Saloner at the Weill Institute for Neurosciences, University of California, San Francisco, together with collaborators, used aptamer-based proteomics to perform in-depth analysis of cerebrospinal fluid samples from 116 individuals carrying FTLD-related pathogenic mutations (C9orf72, GRN, MAPT) and 39 non-carriers, measuring more than 4,000 proteins and constructing 31 protein co-expression modules. The study found that several protein modules were highly correlated with the clinical phenotype of FTLD: RNA splicing-related modules were significantly up-regulated in the C9orf72 and GRN mutation groups, while extracellular matrix modules were specifically elevated in the MAPT mutation group, and both were closely associated with clinical disease severity. Synaptic/neuronal function modules and autophagy-related modules were generally down-regulated, pointing to impaired neuronal function and disturbed protein homeostasis. In addition, this genetic FTLD proteomic signature was successfully validated in two independent cohorts (sporadic progressive supranuclear palsy–Richardson syndrome and the frontotemporal dementia spectrum syndrome), indicating that it has good biological stability and broad applicability.

Pre-analytical bias factors in magnetic bead-enriched plasma proteomics

Enrichment technologies based on nanomaterials have become an important means of breaking through the dynamic range limits of plasma proteomics and extending the depth of detection. Their stability under different sample conditions, however, still lacks systematic validation. The teams of Philipp Emmanuel Geyer and Matthias Mann at the Max Planck Institute of Biochemistry in Germany systematically evaluated five plasma proteomics workflows — three magnetic bead enrichment methods, perchloric acid precipitation (PCA-N) and conventional direct plasma analysis — for their ability to detect low-abundance proteins and their sensitivity to blood cell contamination, by adding low-abundance protein standards and establishing cell contamination models. The study found that although the magnetic bead method improved the detection rate of low-abundance proteins, it was extremely sensitive to contamination by platelets and peripheral blood mononuclear cells (PBMC): trace contamination was enough to produce a falsely inflated count of protein identifications. By contrast, the PCA-N method showed strong tolerance to red blood cell contamination. Through further work with exogenous yeast protein and cell contamination models, as well as different blood collection and centrifugation conditions, the teams proposed quantitative contamination assessment strategies and recommendations for optimising sample handling, underlining the need to strengthen sample quality control in clinical proteomic research and biomarker discovery.

Non-invasive prognostic classification of intratumoural heterogeneity in liver cancer, with multi-omics characterisation and therapeutic implications

Intratumoural heterogeneity (ITH) in hepatocellular carcinoma (HCC) is an important cause of treatment failure and disease recurrence. Conventional biopsy, however, makes sample acquisition difficult, is costly and carries a high risk of complications, and is also vulnerable to sampling bias, making it hard to assess ITH accurately. How to characterise ITH in a non-invasive and comprehensive way remains a major challenge. The team of Professor Liang Xiao at the Zhejiang Provincial Key Laboratory of Precision Diagnosis and Treatment of Liver Diseases, Sir Run Run Shaw Hospital, Zhejiang University School of Medicine, has for the first time built an MRI-based non-invasive prognostic classification system for HCC ITH (RITH), and through multi-omics analysis revealed its molecular mechanisms, offering an entirely new direction for precision diagnosis and treatment of liver cancer. The study collected data from 851 patients with liver cancer at five centres, extracted ITH-related radiomics features from multi-sequence MRI, and constructed and validated the RITH phenotype; transcriptomic and metabolomic data were combined to clarify its molecular mechanism; and potential treatment strategies were further screened on the basis of RITH classification and validated in external and TCGA-TCIA datasets. The results showed that the extracellular matrix–receptor interaction pathway can serve as a potential target for tumours with high RITH. By fusing radiomics with multi-omics, this study achieves fine non-invasive classification of ITH and lays the foundation for developing individualised treatment strategies for liver cancer. Given the widespread use of imaging in tumour diagnosis and treatment, this method also provides a general framework for multi-omics integration analysis of other solid tumours.

The teams of Jared O. Kafader and Neil L. Kelleher at Northwestern University, USA, have developed and applied a new single-cell protein variant imaging mass spectrometry technology (scPiMS), achieving high-throughput, label-free detection of intact protein variants in 10,809 single rat hippocampal cells and overcoming the current technical bottlenecks of single-cell proteomics in throughput and depth of protein characterisation. Built on an Orbitrap platform and liquid sampling technology, the method detected an average of about 4,798 protein variant ions per cell, and with the aid of the PAScore algorithm accurately matched 1,289 of these ions to 165 known protein forms. Combining self-developed data analysis tools and pathway enrichment methods, the team successfully classified 93% of cells as neurons (72%), astrocytes (15%) and microglia (6%). Further analysis showed that the expression features and pathway activities of protein variants can effectively distinguish cell types, confirming their important value for resolving cellular heterogeneity.

The team of Connie R. Jimenez at Vrije Universiteit Amsterdam, together with the team of Guo Tiannan at the School of Medicine, Westlake University, has built a pan-cancer proteomic atlas (TPCPA) covering 22 cancer types based on high-throughput DIA-MS technology. The atlas systematically characterises protein expression features across different cancer types as well as within the same cancer type, identifying a series of pan-cancer and cancer-type-specific protein markers, potential therapeutic targets and subtype classifiers. The study shows promising prospects for clinical application in the protein-level resolution of colorectal cancer CMS subtypes and immune features, and shares its data resources comprehensively through an open online platform. The study not only closes the knowledge gap between the transcriptome and protein expression, it also deepens understanding of cancer heterogeneity from a new molecular perspective, providing support for target discovery and the development of protein-level cancer classification tools. At the same time, the high throughput and good reproducibility of DIA-MS in FFPE samples also lay a solid technical foundation for the clinical translation of biomarkers.