Frontiers in bioinformatics最新文献

Introduction: Disease heterogeneity is the hallmark of breast cancer, which is the most common female malignancy. With a disturbing increase in mortality and disease burden, there remains a need for effective early-stage theragnostic and prognostic biomarkers. In this work, we improved on BrcaDx (https://apalania.shinyapps.io/brcadx/) for cancer vs control screening and examined a cluster of adjoining learning problems in breast cancer heterogeneity: (i) identification of metastatic cancers; (ii) molecular subtyping (TNBC, HER2, or luminal); and (iii) histological subtyping (invasive ductal or invasive lobular).

Methods: We analyzed the transcriptomic profiles of breast cancer patients from public-domain databases such as the TCGA using stage-encoded problem-specific statistical models of gene expression and unveiled stage-salient and progression-significant genes. Using a consensus approach, we identified potential machine learning features, and considered six model classes for each learning problem, with hyperparameter optimization on a training dataset and evaluation on a holdout test dataset. A nested approach enabled us to identify the best model class for each learning problem.

Results: External validation of the best models yielded balanced accuracies of 97.42% for cancer vs normal; 88.22% for metastatic v/s non metastatic; 88.79% for ternary molecular subtyping; and ensemble accuracy of 94.23% for histological subtyping. The model for molecular subtyping was validated on a 26-sample TNBC-only out-of-distribution cohort, yielding 25 correct predictions. We performed a late integration of multi-omics datasets by validating the feature space used in each problem with miRNA profiles, methylation profiles, and commercial breast cancer panels.

Discussion: Pending prospective studies, we have translated the models into BC-Predict that forks the best models developed for each problem in a unified interface and provides a complete readout for input instances of expression data, including uncertainty estimates. BC-Predict is freely available for non-commercial purposes at: https://apalania.shinyapps.io/BC-Predict.

Background: This study aims to investigate potential biomarkers associated with NSCLC-BM and elucidate their regulatory roles in critical pathways involved in cerebral metastatic dissemination.

Methods: The identified DEGs were subjected to functional enrichment analysis. PPI networks were predicted using the STRING database and visualized with Cytoscape. Hub genes were subsequently screened from the PPI network to construct a transcription TF-miRNA regulatory network. Subsequent analyses included: survival analysis, immune infiltration assessment and comprehensive mutational profiling.

Results: Among the 56 identified DEGs, 19 were upregulated while 37 were downregulated. GOntology enrichment analysis revealed significant enrichment in immune response, signaling receptor binding, and extracellular region. KEGG pathway analysis demonstrated predominant involvement in cytokine-cytokine receptor interaction and chemokine signaling pathway. Through Cytoscape-based screening, we identified 10 hub genes: CD19, CD27, IL7R, SELL, CCL5, CCR5, PRF1, GZMK, GZMA, and TIGIT. The TF-miRNA regulatory network analysis uncovered 6 transcription factors (STAT5A/B, NFKB1, EGR1, RELA, and CTCF) and 4 miRNAs(hsa-miR-204, hsa-miR-148b, hsa-miR-618, and hsa-miR-103) as critical transcriptional and post-transcriptional regulators of DEGs.Integrated analyses including Kaplan-Meier survival curves, immune infiltration profiling, and comprehensive mutational analysis demonstrated significant associations with brain metastatic progression in the studied cohort.

Conclusion: This study provides novel biomarkers from a unique perspective for the diagnosis, prognosis, and development of molecular-targeted therapies or immunotherapies for brain metastasis in NSCLC.

Introduction: Chemotherapy response variability in cancer patients necessitates novel strategies targeting chemoresistant populations. While combinatorial regimens show promise through synergistic pharmacological interactions, traditional pathway enrichment methods relying on static gene sets fail to capture drug-induced dynamic transcriptional perturbations.

Methods: To address this challenge, we developed the Pathway-Responsive Gene Sets (PRGS) framework to systematically identify chemoresistance-associated pathways and guide therapeutic intervention. Comparative evaluation of three computational strategies (GSEA-like method, Hypergeometric test-based method, Bates test-based method) revealed that the GSEA-like methodology exhibited superior performance, enabling precise identification of drug-induced pathway dysregulation.

Results: Key experimental findings demonstrated PRGS's superiority over conventional Pathway Member Gene Sets (PMGS), exhibiting statistical independence (p < 0.0001) and enhanced detection of chemotherapy-driven pathway dysregulation. Application of PRGS to the GDSC dataset identified 8 resistance-associated pathways. Screening of agents targeting these pathways yielded candidates with predicted anti-resistance activity. An in vitro cellular experiment demonstrated that the bortezomib-bleomycin combination exhibited synergistic cytotoxicity (IDAcomboScore = 0.014) in T47D cells, highlighting the potential of PRGS-guided therapeutic strategies.

Discussion: This study establishes a PRGS-based methodological framework that integrates genomic perturbations with precision oncology, demonstrating its capacity to decode resistance mechanisms and guide therapeutic development through dynamic pathway analysis.

Background: Macrophages play a critical role in chronic rhinosinusitis with nasal polyps (CRSwNP), and their functional imbalance may cause metabolic disturbances. However, the mechanisms of their role in CRSwNP remain unclear. This study aimed to identify CRSwNP biomarkers related to macrophage polarization and metabolism, and elucidate their molecular regulatory mechanisms.

Methods: In this study, transcriptomic data of chronic rhinosinusitis with nasal polyps (CRSwNP) were obtained from public databases. Differentially expressed genes (DEGs) were screened via differential expression analysis. Subsequently, weighted gene co-expression network analysis (WGCNA) was used to identify key module genes related to macrophage polarization-related genes (MP-RGs), which were then cross-referenced with metabolism-related genes to screen for candidate genes. After that, two machine learning methods-least absolute shrinkage and selection operator (LASSO) and random forest (RF)-were applied to further screen these candidate genes. Receiver operating characteristic (ROC) curves for the training set and validation set were constructed, and gene expression validation was conducted to finally determine the biomarkers. Finally, reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was used to verify the expression levels of prognostic genes.

Results: ALOX5, HMOX1, and PLA2G7 were identified as biomarkers for CRSwNP, with AUC >0.7 in both training and validation sets, showing strong diagnostic potential. A nomogram, built on these three biomarkers, exhibited superior diagnostic performance. Enrichment analysis suggested that these biomarkers might be implicated in immune pathways. Furthermore, all three biomarkers were found to be correlated with asthma. Selenium was identified as a co-target of ALOX5 and HMOX1, presenting potential therapeutic targets for CRSwNP. A total of 10 key miRNAs regulating these biomarkers were identified, and the upstream long non-coding RNAs of hsa-miR-642a-5p, including FOXC1 and NEAT1, were predicted. Additionally, the transcription factor FOXC1 was found to concurrently regulate all three biomarkers. RT-qPCR results validated that the expression levels of ALOX5, HMOX1, and PLA2G7 were significantly elevated in CRSwNP patients, corroborating the findings from bioinformatics analyses.

Conclusion: ALOX5, HMOX1, and PLA2G7 were identified as biomarkers linked to macrophage polarization and metabolism in CRSwNP. These findings offer new insights for early prevention strategies and clinical drug development in CRSwNP.

Introduction: Alternative splicing (AS) of mRNAs is a highly conserved mechanism which can greatly expand the functional diversity of the transcriptome. Aberrant splicing underpins many diseases, and a better understanding of AS can provide insights regarding the molecular mechanisms involved. Importantly, AS can be affected by genetic variants and several studies have indicated large numbers of splicing quantitative trait loci (sQTL). With the advance of single-cell technology, expression QTL studies have been expanded to identify cell type level variants.

Methods: We collected eight full-length scRNA-seq pancreatic islet datasets. Genotyping for each individual was done by the CTAT pipeline and Streka2. The isoform quantification was done by RSEM. Finally, sQTL was obtained by sQTLseeker2.

Results: As a result, we identified 228 cell type level sQTLs for alpha and beta cells across 152 genes. In particular, our study highlights four variants affecting CDC42, a gene related to cell morphology, which have not been observed from bulk sQTL analysis.

Discussion: Our results provide a proof of concept that it is possible to identify cell type level sQTLs, and we envision that better powered studies will allow us to further uncover the genetic regulation of splicing.

This dataset presents a structure-enriched resource of theoretical and empirical SARS-CoV-2 spike receptor-binding domain (RBD) variants, developed under the STAYAHEAD project for pandemic preparedness. It integrates large-scale in silico structure predictions with empirical biophysical measurements. The dataset includes 3,705 single-point Wuhan-Hu-1 RBD variants and 100 higher-order Omicron BA.1/BA.2 variants, annotated with AlphaFold2 and ESMFold metrics and Bio2Byte sequence-based predictors. Structural descriptors-RMSD, TM-score, plDDT, solvent accessibility, hydrophobicity, aggregation propensity-are linked to ACE2 binding and expression data from deep mutational scanning. Provided as a FAIR² Data Package, it supports structure-function analysis, variant modeling, and responsible reuse in virology, structural biology, and computational protein science. This collaboration was co-funded by the PPP Allowance from Health ∼ Holland, Top Sector Life Sciences and Health, to stimulate public-private partnerships.

Structure-based drug design (SBDD) is enhanced by machine learning (ML) to improve both virtual screening and de novo design. Despite advances in ML tools for both strategies, screening remains bounded by time and computational cost, while generative models frequently produce invalid and synthetically inaccessible molecules. Screening time can be improved with pharmacophore search, which quickly identifies ligands in a database that match a pharmacophore query. In this work, we introduce PharmacoForge, a diffusion model for generating 3D pharmacophores conditioned on a protein pocket. Generated pharmacophore queries identify ligands that are guaranteed to be valid, commercially available molecules. We evaluate PharmacoForge against automated pharmacophore generation methods using the LIT-PCBA benchmark and ligand generative models through a docking-based evaluation framework. We further assess pharmacophore quality through a retrospective screening of the DUD-E dataset. PharmacoForge surpasses other pharmacophore generation methods in the LIT-PCBA benchmark, and resulting ligands from pharmacophore queries performed similarly to de novo generated ligands when docking to DUD-E targets and had lower strain energies compared to de novo generated ligands.

Aggressive cancers, such as pancreatic ductal adenocarcinoma (PDAC), are often characterized by a complex and desmoplastic tumor microenvironment, a stroma rich supportive connective tissue composed primarily of extracellular matrix (ECM) and non-cancerous cells. Desmoplasia, a dense deposition of stroma, is a major reason for therapy resistance, acting both as a physical barrier that interferes with drug penetration and as a supportive niche that protects cancer cells through diverse mechanisms. Precise understanding of spatial cell interactions in stroma-rich tumors is essential for optimizing therapeutic responses. It enables detailed mapping of stromal-tumor interfaces, comprehensive cell phenotyping, and insights into changes in tissue architecture, improving assessment of drug responses. Recent advances in multiplexed immunofluorescence imaging have enabled the acquisition of large batches of whole-slide tumor images, but scalable and reproducible methods to analyze the spatial distribution of cell states relative to stromal regions remain limited. To address this gap, we developed an open-source computational pipeline that integrates QuPath, StarDist, and custom Python scripts to quantify biomarker expression at a single- and sub-cellular resolution across entire tumor sections. Our workflow includes: (i) automated nuclei segmentation using StarDist, (ii) machine learning-based cell classification using multiplexed marker expression, (iii) modeling of stromal regions based on fibronectin staining, (iv) sensitivity analyses on classification thresholds to ensure robustness across heterogeneous datasets, and (v) distance-based quantification of the proximity of each cell to the stromal border. To improve consistency across slides with variable staining intensities, we introduce a statistical strategy that translates classification thresholds by propagating a chosen reference percentile across the distribution of marker-related cell measurement in each image. We apply this approach to quantify spatial patterns of distribution of the phosphorylated form of the N-Myc downregulated gene 1 (NDRG1), a novel DNA repair protein that conveys signals from the ECM to the nucleus to maintain replication fork homeostasis, and a known cell proliferation marker Ki67 in fibronectin-defined stromal regions in PDAC xenografts. The pipeline is applicable for the analysis of markers of interest in stroma-rich tissues and is publicly available.

Introduction: Single-cell multi-omics has transformed T-cell biology by enabling the simultaneous analysis of T-cell receptor (TCR) sequences, transcriptomes, and surface proteins at the resolution of individual cells. These capabilities are critical for identifying antigen-specific T-cells and accelerating the development of TCR-based immunotherapies.

Methods: Here, we introduce TCRscape, an open-source Python 3 tool designed for high-resolution T-cell receptor clonotype discovery and quantification, optimized for BD Rhapsody™ single-cell multi-omics data.

Results: TCRscape integrates full-length TCR sequence data with gene expression profiles and surface protein expression to enable multimodal clustering of αβ and γδ T-cell populations. It also outputs Seurat-compatible matrices, facilitating downstream visualization and analysis in standard single-cell analysis environments.

Discussion: By bridging clonotype detection with immune cell transcriptome, proteome, and antigen specificity profiling, TCRscape supports rapid identification of dominant T-cell clones and their functional phenotypes, offering a powerful resource for immune monitoring and TCR-engineered therapeutic development. TCRscape can be found at https://github.com/Perik-Zavodskii/TCRscape/.

Proteolytic digestion is an essential process in mass spectrometry-based proteomics for converting proteins into peptides, hence crucial for protein identification and quantification. In a typical proteomics experiment, digestion reagents are selected without prior evaluation of their optimality for detecting proteins or peptides of interest, partly due to the lack of comprehensive and user-friendly predictive tools. In this work, we introduce Protein Cleaver, a web-based application that systematically assesses regions of proteins that are likely or unlikely to be identified, along with extensive sequence and structure annotation and visualization features. We showcase practical examples of Protein Cleaver's usability in drug discovery and highlight proteins that are typically difficult to detect using the most common proteolytic enzymes. We evaluate trypsin and chymotrypsin for identifying G-protein-coupled receptors and discover that chymotrypsin produces significantly more identifiable peptides than trypsin. We perform a bulk digestion analysis and assess 36 proteolytic enzymes for their ability to detect most of cysteine-containing peptides in the human proteome. We anticipate Protein Cleaver to be a valuable auxiliary tool for proteomics scientists.