GPGPS: a robust prognostic gene pair signature of glioma ensembling IDH mutation and 1p/19q co-deletion

Current research status of Raman spectroscopy in glioma detection

Glioma is the most common primary tumor of the nervous system. Conventional diagnostic methods for glioma often involve time-consuming or reliance on externally introduced materials. Consequently, there is an urgent need for rapid and reliable diagnostic techniques. Raman spectroscopy has emerged as a promising tool, offering rapid, accurate, and label-free analysis with high sensitivity and specificity in biomedical applications. In this review, the fundamental principles of Raman spectroscopy have been introduced, and then the progress of applying Raman spectroscopy in biomedical studies has been summarized, including the identification and typing of glioma. The challenges encountered in the clinical application of Raman spectroscopy for glioma have been discussed, and the prospects have also been envisioned.

PAGE-based transfer learning from single-cell to bulk sequencing enhances model generalization for sepsis diagnosis

Sepsis, caused by infections, sparks a dangerous bodily response. The transcriptional expression patterns of host responses aid in the diagnosis of sepsis, but the challenge lies in their limited generalization capabilities. To facilitate sepsis diagnosis, we present an updated version of single-cell Pair-wise Analysis of Gene Expression (scPAGE) using transfer learning method, scPAGE2, dedicated to data fusion between single-cell and bulk transcriptome. Compared to scPAGE, the upgrade to scPAGE2 featured ameliorated Differentially Expressed Gene Pairs (DEPs) for pretraining a model in single-cell transcriptome and retrained it using bulk transcriptome data to construct a sepsis diagnostic model, which effectively transferred cell-layer information from single-cell to bulk transcriptome. Seven datasets across three transcriptome platforms and fluorescence-activated cell sorting (FACS) were used for performance validation. The model involved four DEPs, showing robust performance across next-generation sequencing and microarray platforms, surpassing state-of-the-art models with an average AUROC of 0.947 and an average AUPRC of 0.987. Analysis of scRNA-seq data reveals higher cell proportions with JAM3-PIK3AP1 expression in sepsis monocytes, decreased ARG1-CCR7 in B and T cells. Elevated IRF6-HP in sepsis monocytes confirmed by both scRNA-seq and an independent cohort using FACS. Both the superior performance of the model and the in vitro validation of IRF6-HP in monocytes emphasize that scPAGE2 is effective and robust in the construction of sepsis diagnostic model. We additionally applied scPAGE2 to acute myeloid leukemia and demonstrated its superior classification performance. Overall, we provided a strategy to improve the generalizability of classification model that can be adapted to a broad range of clinical prediction scenarios.

Less is more: relative rank is more informative than absolute abundance for compositional NGS data

High-throughput gene expression data have been extensively generated and utilized in biological mechanism investigations, biomarker detection, disease diagnosis and prognosis. These applications encompass not only bulk transcriptome, but also single cell RNA-seq data. However, extracting reliable biological information from transcriptome data remains challenging due to the constrains of Compositional Data Analysis. Current data preprocessing methods, including dataset normalization and batch effect correction, are insufficient to address these issues and improve data quality for downstream analysis. Alternatively, qualification methods focusing on the relative order of gene expression (ROGER) are more informative than the quantification methods that rely on gene expression abundance. The Pairwise Analysis of Gene expression method is an enhancement of ROGER, designed for data integration in either sample space or feature space. In this review, we summarize the methods applied to transcriptome data analysis and discuss their potentials in predicting clinical outcomes.

Deep learning model to discriminate diverse infection types based on pairwise analysis of host gene expression

Accurate detection of pathogens, particularly distinguishing between Gram-positive and Gram-negative bacteria, could improve disease treatment. Host gene expression can capture the immune system's response to infections caused by various pathogens. Here, we present a deep neural network model, bvnGPS2, which incorporates the attention mechanism based on a large-scale integrated host transcriptome dataset to precisely identify Gram-positive and Gram-negative bacterial infections as well as viral infections. We performed analysis of 4,949 blood samples across 40 cohorts from 10 countries using our previously designed omics data integration method, iPAGE, to select discriminant gene pairs and train the bvnGPS2. The performance of the model was evaluated on six independent cohorts comprising 374 samples. Overall, our deep neural network model shows robust capability to accurately identify specific infections, paving the way for precise medicine strategies in infection treatment and potentially also for identifying subtypes of other diseases.

Intra-sample reversed pairs based on differentially ranked genes reveal biosignature for ovarian cancer

Ovarian cancer, a major gynecological malignancy, often remains undetected until advanced stages, necessitating more effective early screening methods. Existing biomarker based on differential genes often suffers from variations in clinical practice. To overcome the limitations of absolute gene expression values including batch effects and biological heterogeneity, we introduced a pairwise biosignature leveraging intra-sample differentially ranked genes (DRGs) and machine learning for ovarian cancer detection across diverse cohorts. We analyzed ten cohorts comprising 872 samples with 796 ovarian cancer and 76 normal. Our method, DRGpair, involves three stages: intra-sample ranking differential analysis, reversed gene pair analysis, and iterative LASSO regression. We identified four DRG pairs, demonstrating superior diagnostic performance compared to current state-of-the-art biomarkers and differentially expressed genes in seven independent cohorts. This rank-based approach not only reduced computational complexity but also enhanced the specificity and effectiveness of biomarkers, revealing DRGs as promising candidates for ovarian cancer detection and offering a scalable model adaptable to varying cohort characteristics.

Protein structuromics: A new method for protein structure-function crosstalk in glioma

Glioma is a type of tumor that starts in the glial cells of the brain or spine. Since the 1800s, when the disease was first named, its survival rates have always been unsatisfactory. Despite great advances in molecular biology and traditional treatment methods, many questions regarding cancer occurrence and the underlying mechanism remain to be answered. In this study, we assessed the protein structural features of 20 oncogenes and 20 anti-oncogenes via protein structure and dynamic analysis methods and 3D structural and systematic analyses of the structure-function relationships of proteins. All of these results directly indicate that unfavorable group proteins show more complex structures than favorable group proteins. As the tumor cell microenvironment changes, the balance of oncogene-related and anti-oncogene-related proteins is disrupted, and most of the structures of the two groups of proteins will be disrupted. However, more unfavorable group proteins will maintain and refold to achieve their correct shape faster and perform their functions more quickly than favorable group proteins, and the former thus support cancer development. We hope that these analyses will help promote mechanistic research and the development of new treatments for glioma.

Histone demethylase KDM4B accelerates the progression of glioblastoma via the epigenetic regulation of MYC stability

BACKGROUND

Glioblastoma (GBM) is the most malignant and invasive human brain tumor. Histone demethylase 4B (KDM4B) is abnormally expressed in GBM, but the molecular mechanisms by which KDM4B affects the malignant tumor progression are not well defined.

METHODS

GBM cell lines and xenograft tumor samples were subjected to quantitative PCR (qPCR), Western blot, immunohistochemical staining (IHC), as well as ubiquitination, immunoprecipitation (IP), and chromatin immunoprecipitation (ChIP) assays to investigate the role of KDM4B in the progression of GBM.

RESULTS

Here, we report that KDM4B is an epigenetic activator of GBM progression. Abnormal expression of KDM4B is correlated with a poor prognosis in GBM patients. In GBM cell lines, KDM4B silencing significantly inhibited cell survival, proliferation, migration, and invasion, indicating that KDM4B is essential for the anchorage-independent growth and tumorigenic activity of GBM cells. Mechanistically, KDM4B silencing led to downregulation of the oncoprotein MYC and suppressed the expression of cell cycle proteins and epithelial-to-mesenchymal transition (EMT)-related proteins. Furthermore, we found that KDM4B regulates MYC stability through the E3 ligase complex SCFFBXL3+CRY2 and epigenetically activates the transcription of CCNB1 by removing the repressive chromatin mark histone H3 lysine 9 trimethylation (H3K9me3). Finally, we provide evidence that KDM4B epigenetically activates the transcription of miR-181d-5p, which enhances MYC stability.

CONCLUSIONS

Our study has uncovered a KDM4B-dependent epigenetic mechanism in the control of tumor progression, providing a rationale for utilizing KDM4B as a promising therapeutic target for the treatment of MYC-amplified GBM.

Tumor biology, immune infiltration and liver function define seven hepatocellular carcinoma subtypes linked to distinct drivers, survival and drug response

BACKGROUND & AIMS

Tumor heterogeneity is jointly determined by the components of the tumor ecosystem (TES) including tumor cells, immune cells, stromal cells, and non-cellular components. We aimed to identify subtypes using TES-related genes and determine subtype specific drivers and treatments for hepatocellular carcinoma (HCC).

METHODS

We collected 68 genesets depicting tumor biology, immune infiltration, and liver function, totaling 2831 genes, and collected mRNA profiles and clinical data for over 6000 tumors from 65 datasets in the GEO, TCGA, ICGC, and several other databases. We designed a three-step clustering pipeline to identify subtypes. The microenvironment, genomic alteration, and drug response differences were systematically compared among subtypes.

RESULTS

Seven subtypes (TES-1/2/3/4/5/6/7) were revealed in 159 tumors from the CHCC-HBV cohort. We constructed a single sample classifier using paired genes (sscpgsTES). TES subtypes were significantly associated with multiple clinical variables including etiology, and survival in 14 of 17 cohorts and the meta-cohort. TES-1 had the poorest prognosis and highest proliferation level. Both TES-2 and TES-7 were immune-enriched, however, TES-2 had a significantly worse prognosis, and hypoxic and immunosuppressive microenvironment. TES-4 had activated Wnt pathway, driven by CTNNB1 mutation. Good prognosis TES-6 exhibited the best differentiation. TES-5 and TES-3 were considered as novel subclasses by comparing with ten previous subtyping systems. TES-5 tumors had high AFP but good overall survival, and ∼45% of them harbored AXIN1 mutation. TES-3 was immune and stromal desert, may be driven by high copy number alteration burden, and had the poorest response to immune checkpoint inhibitor. TES-1 and TES-2 had significantly lower response to transarterial chemoembolization, but they showed significantly higher sensitivity to compound YM-155.

CONCLUSIONS

Tumor ecosystem subtypes expand existing HCC subtyping systems, have distinct drivers, prognosis, and treatment vulnerabilities.

MrGPS: an m6A-related gene pair signature to predict the prognosis and immunological impact of glioma patients

N6-methyladenosine (m6A) RNA methylation is the predominant epigenetic modification for mRNAs that regulates various cancer-related pathways. However, the prognostic significance of m6A modification regulators remains unclear in glioma. By integrating the TCGA lower-grade glioma (LGG) and glioblastoma multiforme (GBM) gene expression data, we demonstrated that both the m6A regulators and m6A-target genes were associated with glioma prognosis and activated various cancer-related pathways. Then, we paired m6A regulators and their target genes as m6A-related gene pairs (MGPs) using the iPAGE algorithm, among which 122 MGPs were significantly reversed in expression between LGG and GBM. Subsequently, we employed LASSO Cox regression analysis to construct an MGP signature (MrGPS) to evaluate glioma prognosis. MrGPS was independently validated in CGGA and GEO glioma cohorts with high accuracy in predicting overall survival. The average area under the receiver operating characteristic curve (AUC) at 1-, 3- and 5-year intervals were 0.752, 0.853 and 0.831, respectively. Combining clinical factors of age and radiotherapy, the AUC of MrGPS was much improved to around 0.90. Furthermore, CIBERSORT and TIDE algorithms revealed that MrGPS is indicative for the immune infiltration level and the response to immune checkpoint inhibitor therapy in glioma patients. In conclusion, our study demonstrated that m6A methylation is a prognostic factor for glioma and the developed prognostic model MrGPS holds potential as a valuable tool for enhancing patient management and facilitating accurate prognosis assessment in cases of glioma.

Co-expression module analysis reveals high expression homogeneity for both coding and non-coding genes in sepsis

Sepsis is a life-threatening condition characterized by a harmful host response to infection with organ dysfunction. Annually about 20 million people are dead owing to sepsis and its mortality rates is as high as 20%. However, no studies have been carried out to investigate sepsis from the system biology point of view, as previous research predominantly focused on individual genes without considering their interactions and associations. Here, we conducted a comprehensive exploration of genome-wide expression alterations in both mRNAs and long non-coding RNAs (lncRNAs) in sepsis, using six microarray datasets. Co-expression networks were conducted to identify mRNA and lncRNA modules, respectively. Comparing these sepsis modules with normal modules, we observed a homogeneous expression pattern within the mRNA/lncRNA members, with the majority of them displaying consistent expression direction. Moreover, we identified consistent modules across diverse datasets, consisting of 20 common mRNA members and two lncRNAs, namely CHRM3-AS2 and PRKCQ-AS1, which are potential regulators of sepsis. Our results reveal that the up-regulated common mRNAs are mainly involved in the processes of neutrophil mediated immunity, while the down-regulated mRNAs and lncRNAs are significantly overrepresented in T-cell mediated immunity functions. This study sheds light on the co-expression patterns of mRNAs and lncRNAs in sepsis, providing a novel perspective and insight into the sepsis transcriptome, which may facilitate the exploration of candidate therapeutic targets and molecular biomarkers for sepsis.

bvnGPS: a generalizable diagnostic model for acute bacterial and viral infection using integrative host transcriptomics and pretrained neural networks

MOTIVATION

The confusion of acute inflammation infected by virus and bacteria or noninfectious inflammation will lead to missing the best therapy occasion resulting in poor prognoses. The diagnostic model based on host gene expression has been widely used to diagnose acute infections, but the clinical usage was hindered by the capability across different samples and cohorts due to the small sample size for signature training and discovery.

RESULTS

Here, we construct a large-scale dataset integrating multiple host transcriptomic data and analyze it using a sophisticated strategy which removes batch effect and extracts the common information from different cohorts based on the relative expression alteration of gene pairs. We assemble 2680 samples across 16 cohorts and separately build gene pair signature (GPS) for bacterial, viral, and noninfected patients. The three GPSs are further assembled into an antibiotic decision model (bacterial-viral-noninfected GPS, bvnGPS) using multiclass neural networks, which is able to determine whether a patient is bacterial infected, viral infected, or noninfected. bvnGPS can distinguish bacterial infection with area under the receiver operating characteristic curve (AUC) of 0.953 (95% confidence interval, 0.948-0.958) and viral infection with AUC of 0.956 (0.951-0.961) in the test set (N = 760). In the validation set (N = 147), bvnGPS also shows strong performance by attaining an AUC of 0.988 (0.978-0.998) on bacterial-versus-other and an AUC of 0.994 (0.984-1.000) on viral-versus-other. bvnGPS has the potential to be used in clinical practice and the proposed procedure provides insight into data integration, feature selection and multiclass classification for host transcriptomics data.

AVAILABILITY AND IMPLEMENTATION

The codes implementing bvnGPS are available at https://github.com/Ritchiegit/bvnGPS. The construction of iPAGE algorithm and the training of neural network was conducted on Python 3.7 with Scikit-learn 0.24.1 and PyTorch 1.7. The visualization of the results was implemented on R 4.2, Python 3.7, and Matplotlib 3.3.4.