VSClust: feature-based variance-sensitive clustering of omics data

MetaboLink: A web application for Streamlined Processing and Analysis of Large-Scale Untargeted Metabolomics Data

MOTIVATION

The post-processing and analysis of large-scale untargeted metabolomics data face significant challenges due to the intricate nature of correction, filtration, imputation, and normalization steps. Manual execution across various applications often leads to inefficiencies, human-induced errors, and inconsistencies within the workflow.

RESULTS

Addressing these issues, we introduce MetaboLink, a novel web application designed to process LC-MS metabolomics datasets combining established methodologies and offering flexibility and ease of implementation. It offers visualization options for data interpretation, an interface for statistical testing, and integration with PolySTest for further tests and with VSClust for clustering analysis.

AVAILABILITY

Fully functional tool is publicly available at https://computproteomics.bmb.sdu.dk/Metabolomics/. The source code is available at https://github.com/anitamnd/MetaboLink and a detailed description of the app can be found at https://github.com/anitamnd/MetaboLink/wiki. A tutorial video can be found at https://youtu.be/ZM6j10S6Z8Q.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Identification of structural and regulatory cell-shape determinants in Haloferax volcanii

Archaea play indispensable roles in global biogeochemical cycles, yet many crucial cellular processes, including cell-shape determination, are poorly understood. Haloferax volcanii, a model haloarchaeon, forms rods and disks, depending on growth conditions. Here, we used a combination of iterative proteomics, genetics, and live-cell imaging to identify mutants that only form rods or disks. We compared the proteomes of the mutants with wild-type cells across growth phases, thereby distinguishing between protein abundance changes specific to cell shape and those related to growth phases. The results identified a diverse set of proteins, including predicted transporters, transducers, signaling components, and transcriptional regulators, as important for cell-shape determination. Through phenotypic characterization of deletion strains, we established that rod-determining factor A (RdfA) and disk-determining factor A (DdfA) are required for the formation of rods and disks, respectively. We also identified structural proteins, including an actin homolog that plays a role in disk-shape morphogenesis, which we named volactin. Using live-cell imaging, we determined volactin's cellular localization and showed its dynamic polymerization and depolymerization. Our results provide insights into archaeal cell-shape determination, with possible implications for understanding the evolution of cell morphology regulation across domains.

Dissecting tumor microenvironment heterogeneity in syngeneic mouse models: insights on cancer-associated fibroblast phenotypes shaped by infiltrating T cells

Murine syngeneic tumor models have been used extensively for cancer research for several decades and have been instrumental in driving the discovery and development of cancer immunotherapies. These tumor models are very simplistic cancer models, but recent reports have, however, indicated that the different inoculated cancer cell lines can lead to the formation of unique tumor microenvironments (TMEs). To gain more knowledge from studies based on syngeneic tumor models, it is essential to obtain an in-depth understanding of the cellular and molecular composition of the TME in the different models. Additionally, other parameters that are important for cancer progression, such as collagen content and mechanical tissue stiffness across syngeneic tumor models have not previously been reported. Here, we compare the TME of tumors derived from six common syngeneic tumor models. Using flow cytometry and transcriptomic analyses, we show that strikingly unique TMEs are formed by the different cancer cell lines. The differences are reflected as changes in abundance and phenotype of myeloid, lymphoid, and stromal cells in the tumors. Gene expression analyses support the different cellular composition of the TMEs and indicate that distinct immunosuppressive mechanisms are employed depending on the tumor model. Cancer-associated fibroblasts (CAFs) also acquire very different phenotypes across the tumor models. These differences include differential expression of genes encoding extracellular matrix (ECM) proteins, matrix metalloproteinases (MMPs), and immunosuppressive factors. The gene expression profiles suggest that CAFs can contribute to the formation of an immunosuppressive TME, and flow cytometry analyses show increased PD-L1 expression by CAFs in the immunogenic tumor models, MC38 and CT26. Comparison with CAF subsets identified in other studies shows that CAFs are skewed towards specific subsets depending on the model. In athymic mice lacking tumor-infiltrating cytotoxic T cells, CAFs express lower levels of PD-L1 and lower levels of fibroblast activation markers. Our data underscores that CAFs can be involved in the formation of an immunosuppressive TME.

Statistical Analysis of Post-Translational Modifications Quantified by Label-Free Proteomics Across Multiple Biological Conditions with R: Illustration from SARS-CoV-2 Infected Cells

Protein post-translational modifications (PTMs) are essential elements of cellular communication. Their variations in abundance can affect cellular pathways, leading to cellular disorders and diseases. A widely used method for revealing PTM-mediated regulatory networks is their label-free quantitation (LFQ) by high-resolution mass spectrometry. The raw data resulting from such experiments are generally interpreted using specific software, such as MaxQuant, MassChroQ, or Proline for instance. They provide data matrices containing quantified intensities for each modified peptide identified. Statistical analyses are then necessary (1) to ensure that the quantified data are of good enough quality and sufficiently reproducible, (2) to highlight the modified peptides that are differentially abundant between the biological conditions under study. The objective of this chapter is therefore to provide a complete data analysis pipeline for analyzing the quantified values of modified peptides in presence of two or more biological conditions using the R software. We illustrate our pipeline starting from MaxQuant outputs dealing with the analysis of A549-ACE2 cells infected by SARS-CoV-2 at different time stamps, freely available on PRIDE (PXD020019).

Proteomic Profile of Procoagulant Extracellular Vesicles Reflects Complement System Activation and Platelet Hyperreactivity of Patients with Severe COVID-19

Background

Extracellular vesicles (EVs) are a valuable source of biomarkers and display the pathophysiological status of various diseases. In COVID-19, EVs have been explored in several studies for their ability to reflect molecular changes caused by SARS-CoV-2. Here we provide insights into the roles of EVs in pathological processes associated with the progression and severity of COVID-19.

Methods

In this study, we used a label-free shotgun proteomic approach to identify and quantify alterations in EV protein abundance in severe COVID-19 patients. We isolated plasma extracellular vesicles from healthy donors and patients with severe COVID-19 by size exclusion chromatography (SEC). Then, flow cytometry was performed to assess the origin of EVs and to investigate the presence of circulating procoagulant EVs in COVID-19 patients. A total protein extraction was performed, and samples were analyzed by nLC-MS/MS in a Q-Exactive HF-X. Finally, computational analysis was applied to signify biological processes related to disease pathogenesis.

Results

We report significant changes in the proteome of EVs from patients with severe COVID-19. Flow cytometry experiments indicated an increase in total circulating EVs and with tissue factor (TF) dependent procoagulant activity. Differentially expressed proteins in the disease groups were associated with complement and coagulation cascades, platelet degranulation, and acute inflammatory response.

Conclusions

The proteomic data reinforce the changes in the proteome of extracellular vesicles from patients infected with SARS-CoV-2 and suggest a role for EVs in severe COVID-19.

Aedes aegypti Aag-2 Cell Proteome Modulation in Response to Chikungunya Virus Infection

Chikungunya virus (CHIKV) is a single-stranded positive RNA virus that belongs to the genus Alphavirus and is transmitted to humans by infected Aedes aegypti and Aedes albopictus bites. In humans, CHIKV usually causes painful symptoms during acute and chronic stages of infection. Conversely, virus–vector interaction does not disturb the mosquito’s fitness, allowing a persistent infection. Herein, we studied CHIKV infection of Ae. aegypti Aag-2 cells (multiplicity of infection (MOI) of 0.1) for 48 h through label-free quantitative proteomic analysis and transmission electron microscopy (TEM). TEM images showed a high load of intracellular viral cargo at 48 h postinfection (hpi), as well as an unusual elongated mitochondria morphology that might indicate a mitochondrial imbalance. Proteome analysis revealed 196 regulated protein groups upon infection, which are related to protein synthesis, energy metabolism, signaling pathways, and apoptosis. These Aag-2 proteins regulated during CHIKV infection might have roles in antiviral and/or proviral mechanisms and the balance between viral propagation and the survival of host cells, possibly leading to the persistent infection.

Diel investments in metabolite production and consumption in a model microbial system

Organic carbon transfer between surface ocean photosynthetic and heterotrophic microbes is a central but poorly understood process in the global carbon cycle. In a model community in which diatom extracellular release of organic molecules sustained growth of a co-cultured bacterium, we determined quantitative changes in the diatom endometabolome and the bacterial uptake transcriptome over two diel cycles. Of the nuclear magnetic resonance (NMR) peaks in the diatom endometabolites, 38% had diel patterns with noon or mid-afternoon maxima; the remaining either increased (36%) or decreased (26%) through time. Of the genes in the bacterial uptake transcriptome, 94% had a diel pattern with a noon maximum; the remaining decreased over time (6%). Eight diatom endometabolites identified with high confidence were matched to the bacterial genes mediating their utilization. Modeling of these coupled inventories with only diffusion-based phytoplankton extracellular release could not reproduce all the patterns. Addition of active release mechanisms for physiological balance and bacterial recognition significantly improved model performance. Estimates of phytoplankton extracellular release range from only a few percent to nearly half of annual net primary production. Improved understanding of the factors that influence metabolite release and consumption by surface ocean microbes will better constrain this globally significant carbon flux.

Multi-level attention graph neural network based on co-expression gene modules for disease diagnosis and prognosis

MOTIVATION

Advanced deep learning techniques have been widely applied in disease diagnosis and prognosis with clinical omics, especially gene expression data. In the regulation of biological processes and disease progression, genes often work interactively rather than individually. Therefore, investigating gene association information and co-functional gene modules can facilitate disease state prediction.

RESULTS

To explore the gene modules and inter-gene relational information contained in the omics data, we propose a novel multi-level attention graph neural network (MLA-GNN) for disease diagnosis and prognosis. Specifically, we format omics data into co-expression graphs via weighted correlation network analysis, and then construct multi-level graph features, finally fuse them through a well-designed multi-level graph feature fully fusion module to conduct predictions. For model interpretation, a novel full-gradient graph saliency mechanism is developed to identify the disease-relevant genes. MLA-GNN achieves state-of-the-art performance on transcriptomic data from TCGA-LGG/TCGA-GBM and proteomic data from coronavirus disease 2019 (COVID-19)/non-COVID-19 patient sera. More importantly, the relevant genes selected by our model are interpretable and are consistent with the clinical understanding.

AVAILABILITYAND IMPLEMENTATION

The codes are available at https://github.com/TencentAILabHealthcare/MLA-GNN.

Impaired glucocorticoid receptor expression in liver disrupts feeding-induced gene expression, glucose uptake, and glycogen storage

The transition from a fasted to a fed state is associated with extensive transcriptional remodeling in hepatocytes facilitated by hormonal- and nutritional-regulated transcription factors. Here, we use a liver-specific glucocorticoid receptor (GR) knockout (L-GRKO) model to investigate the temporal hepatic expression of GR target genes in response to feeding. Interestingly, in addition to the well-described fasting-regulated genes, we identify a subset of hepatic feeding-induced genes that requires GR for full expression. This includes Gck, which is important for hepatic glucose uptake, utilization, and storage. We show that insulin and glucocorticoids cooperatively regulate hepatic Gck expression in a direct GR-dependent manner by a 4.6 kb upstream GR binding site operating as a Gck enhancer. L-GRKO blunts preprandial and early postprandial Gck expression, which ultimately affects early postprandial hepatic glucose uptake, phosphorylation, and glycogen storage. Thus, GR is positively involved in feeding-induced gene expression and important for postprandial glucose metabolism in the liver.

Inflammatory capacity of exosomes released in the early stages of acute pancreatitis predicts the severity of the disease

As acute pancreatitis progresses to the severe form, a life-threatening systemic inflammation is triggered. Although the mechanisms involved in this process are not yet well understood, it has been proposed that circulating exosomes may be involved in the progression of inflammation from the pancreas to distant organs. Here, the inflammatory capacity and protein profile of plasma exosomes obtained during the first 24 h of hospitalization of patients diagnosed with acute pancreatitis were characterized and compared with the final severity of the disease. We found that the final severity of the disease strongly correlates with the inflammatory capacity of exosomes in the early stages of acute pancreatitis. Exosomes isolated from patients with mild pancreatitis had no effect on macrophages, while exosomes isolated from patients with severe pancreatitis triggered NFκB activation, TNFα and IL1β expression, and free radical generation. To delve deeper into the mechanism involved, we performed a proteomic analysis of the different exosomes that allowed us to identify different groups of proteins whose concentration was also correlated with the clinical classification of pancreatitis. In particular, an increase in the amount of S100A8 and S100A9 carried by exosomes of severe pancreatitis suggests that the mechanism of action of exosomes is mediated by the effect of these proteins on NADPH oxidase. This enzyme is activated by S100A8/S100A9, thus generating free radicals and promoting an inflammatory response. Along these lines, we observed that inhibition of this enzyme abolished all the pro-inflammatory effects of exosomes from severe pancreatitis. All this suggests that the systemic effects, and therefore the final severity of acute pancreatitis, are determined by the content of circulating exosomes generated in the early hours of the process. © 2021 The Authors. The Journal of Pathology published by John Wiley & Sons, Ltd. on behalf of The Pathological Society of Great Britain and Ireland.

Super.FELT: supervised feature extraction learning using triplet loss for drug response prediction with multi-omics data

BACKGROUND

Predicting the drug response of a patient is important for precision oncology. In recent studies, multi-omics data have been used to improve the prediction accuracy of drug response. Although multi-omics data are good resources for drug response prediction, the large dimension of data tends to hinder performance improvement. In this study, we aimed to develop a new method, which can effectively reduce the large dimension of data, based on the supervised deep learning model for predicting drug response.

RESULTS

We proposed a novel method called Supervised Feature Extraction Learning using Triplet loss (Super.FELT) for drug response prediction. Super.FELT consists of three stages, namely, feature selection, feature encoding using a supervised method, and binary classification of drug response (sensitive or resistant). We used multi-omics data including mutation, copy number aberration, and gene expression, and these were obtained from cell lines [Genomics of Drug Sensitivity in Cancer (GDSC), Cancer Cell Line Encyclopedia (CCLE), and Cancer Therapeutics Response Portal (CTRP)], patient-derived tumor xenografts (PDX), and The Cancer Genome Atlas (TCGA). GDSC was used for training and cross-validation tests, and CCLE, CTRP, PDX, and TCGA were used for external validation. We performed ablation studies for the three stages and verified that the use of multi-omics data guarantees better performance of drug response prediction. Our results verified that Super.FELT outperformed the other methods at external validation on PDX and TCGA and was good at cross-validation on GDSC and external validation on CCLE and CTRP. In addition, through our experiments, we confirmed that using multi-omics data is useful for external non-cell line data.

CONCLUSION

By separating the three stages, Super.FELT achieved better performance than the other methods. Through our results, we found that it is important to train encoders and a classifier independently, especially for external test on PDX and TCGA. Moreover, although gene expression is the most powerful data on cell line data, multi-omics promises better performance for external validation on non-cell line data than gene expression data. Source codes of Super.FELT are available at  https://github.com/DMCB-GIST/Super.FELT .

A Tutorial for Variance-Sensitive Clustering and the Quantitative Analysis of Protein Complexes

Data clustering facilitates the identification of biologically relevant molecular features in quantitative proteomics experiments with thousands of measurements over multiple conditions. It finds groups of proteins or peptides with similar quantitative behavior across multiple experimental conditions. This co-regulatory behavior suggests that the proteins of such a group share their functional behavior and thus often can be mapped to the same biological processes and molecular subnetworks.While usual clustering approaches dismiss the variance of the measured proteins, VSClust combines statistical testing with pattern recognition into a common algorithm. Here, we show how to use the VSClust web service on a large proteomics data set and present further tools to assess the quantitative behavior of protein complexes.

Proteomic Studies of Primary Acute Myeloid Leukemia Cells Derived from Patients Before and during Disease-Stabilizing Treatment Based on All-Trans Retinoic Acid and Valproic Acid

All-trans retinoic acid (ATRA) and valproic acid (VP) have been tried in the treatment of non-promyelocytic variants of acute myeloid leukemia (AML). Non-randomized studies suggest that the two drugs can stabilize AML and improve normal peripheral blood cell counts. In this context, we used a proteomic/phosphoproteomic strategy to investigate the in vivo effects of ATRA/VP on human AML cells. Before starting the combined treatment, AML responders showed increased levels of several proteins, especially those involved in neutrophil degranulation/differentiation, M phase regulation and the interconversion of nucleotide di- and triphosphates (i.e., DNA synthesis and binding). Several among the differentially regulated phosphorylation sites reflected differences in the regulation of RNA metabolism and apoptotic events at the same time point. These effects were mainly caused by increased cyclin dependent kinase 1 and 2 (CDK1/2), LIM domain kinase 1 and 2 (LIMK1/2), mitogen-activated protein kinase 7 (MAPK7) and protein kinase C delta (PRKCD) activity in responder cells. An extensive effect of in vivo treatment with ATRA/VP was the altered level and phosphorylation of proteins involved in the regulation of transcription/translation/RNA metabolism, especially in non-responders, but the regulation of cell metabolism, immune system and cytoskeletal functions were also affected. Our analysis of serial samples during the first week of treatment suggest that proteomic and phosphoproteomic profiling can be used for the early identification of responders to ATRA/VP-based treatment.

Lipid molecular timeline profiling reveals diurnal crosstalk between the liver and circulation

Diurnal regulation of whole-body lipid metabolism plays a vital role in metabolic health. Although changes in lipid levels across the diurnal cycle have been investigated, the system-wide molecular responses to both short-acting fasting-feeding transitions and longer-timescale circadian rhythms have not been explored in parallel. Here, we perform time-series multi-omics analyses of liver and plasma revealing that the majority of molecular oscillations are entrained by adaptations to fasting, food intake, and the postprandial state. By developing algorithms for lipid structure enrichment analysis and lipid molecular crosstalk between tissues, we find that the hepatic phosphatidylethanolamine (PE) methylation pathway is diurnally regulated, giving rise to two pools of oscillating phosphatidylcholine (PC) molecules in the circulation, which are coupled to secretion of either very low-density lipoprotein (VLDL) or high-density lipoprotein (HDL) particles. Our work demonstrates that lipid molecular timeline profiling across tissues is key to disentangling complex metabolic processes and provides a critical resource for the study of whole-body lipid metabolism.

Biological characteristics of aging in human acute myeloid leukemia cells: the possible importance of aldehyde dehydrogenase, the cytoskeleton and altered transcriptional regulation

Patients with acute myeloid leukemia (AML) have a median age of 65-70 years at diagnosis. Elderly patients have more chemoresistant disease, and this is partly due to decreased frequencies of favorable and increased frequencies of adverse genetic abnormalities. However, aging-dependent differences may also contribute. We therefore compared AML cell proteomic and phosphoproteomic profiles for (i) elderly low-risk and younger low-risk patients with favorable genetic abnormalities; and (ii) high-risk patients with adverse genetic abnormalities and a higher median age against all low-risk patients with lower median age. Elderly low-risk and younger low-risk patients showed mainly phosphoproteomic differences especially involving transcriptional regulators and cytoskeleton. When comparing high-risk and low-risk patients both proteomic and phosphoproteomic studies showed differences involving cytoskeleton and immunoregulation but also transcriptional regulation and cell division. The age-associated prognostic impact of cyclin-dependent kinases was dependent on the cellular context. The protein level of the adverse prognostic biomarker mitochondrial aldehyde dehydrogenase (ALDH2) showed a similar significant upregulation both in elderly low-risk and elderly high-risk patients. Our results suggest that molecular mechanisms associated with cellular aging influence chemoresistance of AML cells, and especially the cytoskeleton function may then influence cellular hallmarks of aging, e.g. mitosis, polarity, intracellular transport and adhesion.

Changes in the Oligodendrocyte Progenitor Cell Proteome with Ageing

Following central nervous system (CNS) demyelination, adult oligodendrocyte progenitor cells (OPCs) can differentiate into new myelin-forming oligodendrocytes in a regenerative process called remyelination. Although remyelination is very efficient in young adults, its efficiency declines progressively with ageing. Here we performed proteomic analysis of OPCs freshly isolated from the brains of neonate, young and aged female rats. Approximately 50% of the proteins are expressed at different levels in OPCs from neonates compared with their adult counterparts. The amount of myelin-associated proteins, and proteins associated with oxidative phosphorylation, inflammatory responses and actin cytoskeletal organization increased with age, whereas cholesterol-biosynthesis, transcription factors and cell cycle proteins decreased. Our experiments provide the first ageing OPC proteome, revealing the distinct features of OPCs at different ages. These studies provide new insights into why remyelination efficiency declines with ageing and potential roles for aged OPCs in other neurodegenerative diseases.

Dynamic proteomic analysis of Aedes aegypti Aag-2 cells infected with Mayaro virus

Background

Mayaro virus (MAYV) is responsible for a mosquito-borne tropical disease with clinical symptoms similar to dengue or chikungunya virus fevers. In addition to the recent territorial expansion of MAYV, this virus may be responsible for an increasing number of outbreaks. Currently, no vaccine is available. Aedes aegypti is promiscuous in its viral transmission and thus an interesting model to understand MAYV-vector interactions. While the life-cycle of MAYV is known, the mechanisms by which this arbovirus affects mosquito host cells are not clearly understood.

Methods

After defining the best conditions for cell culture harvesting using the highest virus titer, Ae. aegypti Aag-2 cells were infected with a Brazilian MAYV isolate at a MOI of 1 in order to perform a comparative proteomic analysis of MAYV-infected Aag-2 cells by using a label-free semi-quantitative bottom-up proteomic analysis. Time-course analyses were performed at 12 and 48 h post-infection (hpi). After spectrum alignment between the triplicates of each time point and changes of the relative abundance level calculation, the identified proteins were annotated and using Gene Ontology database and protein pathways were annotated using the Kyoto Encyclopedia of Genes and Genomes.

Results

After three reproducible biological replicates, the total proteome analysis allowed for the identification of 5330 peptides and the mapping of 459, 376 and 251 protein groups, at time 0, 12 hpi and 48 hpi, respectively. A total of 161 mosquito proteins were found to be differentially abundant during the time-course, mostly related to host cell processes, including redox metabolism, translation, energy metabolism, and host cell defense. MAYV infection also increased host protein expression implicated in viral replication.

Conclusions

To our knowledge, this first proteomic time-course analysis of MAYV-infected mosquito cells sheds light on the molecular basis of the viral infection process and host cell response during the first 48 hpi. Our data highlight several mosquito proteins modulated by the virus, revealing that MAYV manipulates mosquito cell metabolism for its propagation.

NCutYX: a package for clustering analysis of multilayer omics data

SUMMARY

Multilayer omics profiling has become a major venue for understanding complex diseases. We develop NCutYX, an R package for clustering analysis of multilayer omics data. The package and methods jointly analyze multiple layers of omics measurements and effectively accommodate their regulations. They systematically conduct a series of analysis based on the normalized cut technique, including the clusterings of subjects and omics measurements and biclustering. The package can be valuable for its timely context, novel methods, and comprehensiveness.

AVAILABILITY

https://cran.r-project.org/web/packages/NCutYX/.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

RAFTS3G: an efficient and versatile clustering software to analyses in large protein datasets

BACKGROUND

Clustering methods are essential to partitioning biological samples being useful to minimize the information complexity in large datasets. Tools in this context usually generates data with greed algorithms that solves some Data Mining difficulties which can degrade biological relevant information during the clustering process. The lack of standardization of metrics and consistent bases also raises questions about the clustering efficiency of some methods. Benchmarks are needed to explore the full potential of clustering methods - in which alignment-free methods stand out - and the good choice of dataset makes it essentials.

RESULTS

Here we present a new approach to Data Mining in large protein sequences datasets, the Rapid Alignment Free Tool for Sequences Similarity Search to Groups (RAFTS3G), a method to clustering aiming of losing less biological information in the processes of generation groups. The strategy developed in our algorithm is optimized to be more astringent which reflects increase in accuracy and sensitivity in the generation of clusters in a wide range of similarity. RAFTS3G is the better choice compared to three main methods when the user wants more reliable result even ignoring the ideal threshold to clustering.

CONCLUSION

In general, RAFTS3G is able to group up to millions of biological sequences into large datasets, which is a remarkable option of efficiency in clustering. RAFTS3G compared to other "standard-gold" methods in the clustering of large biological data maintains the balance between the reduction of biological information redundancy and the creation of consistent groups. We bring the binary search concept applied to grouped sequences which shows maintaining sensitivity/accuracy relation and up to minimize the time of data generated with RAFTS3G process.

Coordination between TGF-β cellular signaling and epigenetic regulation during epithelial to mesenchymal transition

BACKGROUND

Epithelial to mesenchymal transition (EMT) plays a crucial role in cancer propagation. It can be orchestrated by the activation of multiple signaling pathways, which have been found to be highly coordinated with many epigenetic regulators. Although the mechanism of EMT has been studied over decades, cross talk between signaling and epigenetic regulation is not fully understood.

RESULTS

Here, we present a time-resolved multi-omics strategy, which featured the identification of the correlation between protein changes (proteome), signaling pathways (phosphoproteome) and chromatin modulation (histone modifications) dynamics during TGF-β-induced EMT. Our data revealed that Erk signaling was activated in 5-min stimulation and structural proteins involved in cytoskeleton rearrangement were regulated after 1-day treatment, constituting a detailed map of systematic changes. The comprehensive profiling of histone post-translational modifications identified H3K27me3 as the most significantly up-regulated mark. We thus speculated and confirmed that a combined inhibition of Erk signaling and Ezh2 (H3K27me3 methyltransferase) was more effective in blocking EMT progress than individual inhibitions.

CONCLUSIONS

In summary, our data provided a more detailed map of cross talk between signaling pathway and chromatin regulation comparing to previous EMT studies. Our findings point to a promising therapeutic strategy for EMT-related diseases by combining Erk inhibitor (singling pathway) and Ezh2 inhibitor (epigenetic regulation).