SEARCH · Search bioRxiv

Results for “Cancer Biology”

Search indexed bioRxiv preprints in genomics, neuroscience, cell biology and bioinformatics. Read source abstracts and check manuscript versions; preprints are not peer reviewed.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 199 records · Page 11Linked to original sources

bioRxiv2025

Pan-Cancer Single-Cell Profiling Uncovers the Biological Characteristics of Cancer-Testis Genes

Cancer-testis genes (CTGs) are attractive immunotherapeutic targets owing to their restricted testicular expression and aberrant activation in cancers. However, their regulatory mechanisms, spatial organization, and clinical utility remain incompletely understood. Here, we leveraged large-scale single-cell and spatial transcriptomic data to perform a comprehensive pan-cancer analysis of CTGs. We established a high-confidence pan-cancer CTG catalog and uncovered a heterogeneous epigenetic regulatory landscape in which X-linked CTGs are predominantly governed by DNA methylation, whereas autosomal CTGs are more strongly associated with chromatin regulators. Building on the observation that CTG activation is a robust pan-cancer hallmark of malignancy, we developed a computational framework that enables rapid malignant cell annotation with performance comparable to established copy number variation-based methods. Spatial transcriptomic analyses revealed that CTG expression in head and neck squamous cell carcinoma is preferentially enriched at the invasive tumor front. Clinically, we highlighted the underappreciated therapeutic potential of CTGs and prioritized CT83 and DCAF4L2 as promising candidate targets for T-cell receptor-engineered T-cell therapy in triple-negative breast cancer and liver cancer, respectively. Our study advances mechanistic understanding of CTG biology and provides a valuable resource for the development of CTG-based immunotherapies.

cancer biology↗

bioRxiv2023

Proteasome inhibition reprograms chromatin landscape in breast cancer

The 26S proteasome is the major protein degradation machinery in cells. Cancer cells use the proteasome to modulate gene expression networks that promote tumor growth. Proteasome inhibitors have emerged as effective cancer therapeutics, but how they work mechanistically remains unclear. Here, using integrative genomic analysis, we discovered unexpected reprogramming of the chromatin landscape and RNAPII transcription initiation in breast cancer cells treated with the proteasome inhibitor MG132. The cells acquired dynamic changes in chromatin accessibility at specific genomic loci termed Differentially Open Chromatin Regions (DOCRs). DOCRs with decreased accessibility were promoter proximal and exhibited unique chromatin architecture associated with divergent RNAPII transcription. Conversely, DOCRs with increased accessibility were primarily distal to transcription start sites and enriched in oncogenic super enhancers predominantly accessible in non-basal breast tumor subtypes. These findings describe the mechanisms by which the proteasome modulates the expression of gene networks intrinsic to breast cancer biology. HighlightsO_LIProteasome inhibition uncovers de novo Differential Open Chromatin Regions (DOCRs) in breast cancer cells. C_LIO_LIProteasome inhibitor sensitive promoters exhibit a distinctive chromatin architecture with discrete transcription initiation patterns. C_LIO_LIProteasome inhibition reprograms accessibility of super enhancers. C_LIO_LIProteasome inhibitor sensitive super enhancers distinguish basal from non-basal breast cancer subtypes. C_LI

genomics↗

bioRxiv2017

polyCluster: Defining Communities of Reconciled Cancer Subtypes with Biological and Prognostic Significance

To stratify cancer patients for most beneficial therapies, it is a priority to define robust molecular subtypes using clustering methods and \"big data\". If each of these methods produces different numbers of clusters for the same data, it is difficult to achieve an optimal solution. Here, we introduce \"polyCluster\", a tool that reconciles clusters identified by different methods into context-specific subtype \"communities\" using a hypergeometric test or a measure of relative proportion of common samples. The polycluster was tested using a breast cancer dataset, and latter using uveal melanoma datasets to identify novel subtype communities with significant metastasis-free prognostic differences. Available at: https://github.com/syspremed/polyClustR

genomics↗

bioRxiv2021

Replication Study: RAF inhibitors prime wild-type RAF to activate the MAPK pathway and enhance growth

As part of the Reproducibility Project: Cancer Biology, we published a Registered Report (Bhargava et al., 2016) that described how we intended to replicate selected experiments from the paper "RAF inhibitors prime wild-type RAF to activate the MAPK pathway and enhance growth" (Hatzivassiliou et al., 2010). Here we report the results. We found two unrelated RAF inhibitors, PLX4720 or GDC-0879, selectively inhibited BRAF(V600E) cell proliferation, while the MEK inhibitor, PD0325901, inhibited BRAF(V600E), wild-type RAF/RAS, and mutant RAS cancer cell proliferation, similar to the original study (Figure 1A; Hatzivassiliou et al., 2010). We found knockdown of CRAF, but not BRAF, in mutant RAS cells attenuated the phospho-MEK induction observed after PLX4720 treatment, similar to the original study (Figure 2B; Hatzivassiliou et al., 2010). The original study reported analogous results with GDC-0879, which was not observed in this replication, although unexpected control results confound the interpretation. We also attempted a replication of an assay with recombinant proteins to test the differential effect of RAF inhibitors on BRAF-CRAF heterodimerization (Figure 4A; Hatzivassiliou et al., 2010). Although we were unable to conduct the experiment as planned, we observed differential binding of BRAF by RAF inhibitors; however, it was between BRAF and beads, independent of CRAF. While these data were unable to address whether, under the conditions of the original study, the same observations could be observed, we discuss key differences between the original study and this replication that are important to consider for further experiments. Finally, where possible, we report meta-analyses for each result. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=62 SRC="FIGDIR/small/470372v2_fig1.gif" ALT="Figure 1"> View larger version (7K): org.highwire.dtl.DTLVardef@1f77d0eorg.highwire.dtl.DTLVardef@1333494org.highwire.dtl.DTLVardef@c10824org.highwire.dtl.DTLVardef@fda30f_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 1.C_FLOATNO Cellular dose response curves for RAF and MEK inhibitors in BRAF(V600E), mutant RAS, and wild-type RAF/RAS cell lines Cell viability assays were performed for RAF inhibitors (PLX4720 and GDC-0879) and MEK inhibitor (PD0325901) against A375, (BRAF(V600E)), HCT116 (mutant RAS), and MeWo (wild-type RAF/RAS) cells. Absolute half-maximum effective concentration (EC50) values ({micro}M) were determined for each biological repeat. EC50 values unable to be accurately estimated are reported as either >20 {micro}M or <0.078 {micro}M, which are the largest and smallest doses tested, respectively. Absolute EC50 values for each biological repeat for A375, HCT116, and MeWo cells treated with (A) PLX4720, (B) GDC-0879, or (C) PD0325901 for this replication attempt are plotted with the EC50 value reported in Hatzivassiliou et al. (2010) displayed as a single point (red triangle) for comparison. Where possible the mean and 95% confidence interval of the replication data are shown. Additional details for this experiment can be found at https://osf.io/52zp9/. C_FIG O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=198 SRC="FIGDIR/small/470372v2_fig2.gif" ALT="Figure 2"> View larger version (34K): org.highwire.dtl.DTLVardef@1dc3058org.highwire.dtl.DTLVardef@e9f5d9org.highwire.dtl.DTLVardef@baf5eorg.highwire.dtl.DTLVardef@126717a_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 2.C_FLOATNO Cellular dose response curves for each biological repeat This is the same experiment as in Figure 1. The dose response curve of each biological repeat [n=4] for A375, HCT116, and MeWo cells treated with (A) PLX4720, (B) GDC-0879, or (C) PD0325901 for this replication attempt are plotted. Additional details for this experiment can be found at https://osf.io/52zp9/. C_FIG O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=111 SRC="FIGDIR/small/470372v2_fig4.gif" ALT="Figure 4"> View larger version (33K): org.highwire.dtl.DTLVardef@e4357eorg.highwire.dtl.DTLVardef@73e9ceorg.highwire.dtl.DTLVardef@1fb01c6org.highwire.dtl.DTLVardef@31a799_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 4.C_FLOATNO Testing of BRAF and CRAF antibodies used in original study The BRAF and CRAF antibodies reported in the original study (Hatzivassiliou et al., 2010) were tested for suitability by Western blot. (A) Representative Western blots and Ponceau-stained membrane (experiment performed independently 3 times) of 0.5 {micro}g of purified recombinant BRAF kinase domain (KD). Membranes were probed with the BRAF antibody used in the original study (RRID:AB_1840330) or with a different BRAF antibody used in this replication attempt (RRID:AB_2721113). (B) Representative Western blots and Ponceau-stained membrane (experiment performed independently 3 times) of 0.5 {micro}g of purified recombinant GST tagged CRAF KD. Membranes were probed with the CRAF antibody used in the original study (RRID:AB_397552) or with a different CRAF antibody used in this replication attempt (RRID:AB_10989977). (C) Representative Western blots (experiment performed once) of 16 {micro}g of whole cell lysates from the indicated human cell lines. Membranes were probed with the BRAF (RRID:AB_1840330) or CRAF (RRID:AB_397552) antibodies used in the original study. (D) Screenshot from the Constraint-based Multiple Alignment Tool (COBALT) (Papadopoulos and Agarwala, 2007) showing the alignment of full-length BRAF (NCBI Reference Sequence: NP_004324.2), the region of BRAF used as the immunogenic peptide used to generate the BRAF antibody used in the original study (RRID:AB_1840330), the BRAF KD used in this replication attempt, and the BRAF KD used in the original study. (E) Screenshot from COBALT showing the alignment of full-length CRAF (NCBI Reference Sequence: NP_002871.1), the region of CRAF used as the immunogenic peptide used to generate the CRAF antibody used in the original study (RRID:AB_397552), and the CRAF KD used in the original study and this replication attempt. Additional details for this experiment can be found at https://osf.io/6f3sk/. C_FIG

cancer biology↗

bioRxiv2019

EuRBPDB: a comprehensive resource for annotation, functional and oncological investigation of eukaryotic RNA binding proteins (RBPs)

RNA binding proteins (RBPs) are a large protein family that plays important roles at almost all levels of gene regulation through interacting with RNAs, and contributes to numerous biological processes. However, the complete list of eukaryotic RBPs including human is still unavailable. In this study, we systematically identified RBPs in 162 eukaryotic species based on both computational analysis of RNA binding domains (RBDs) and large-scale RNA binding proteomic (RBPome) data, and established a comprehensive eukaryotic RBP database, EuRBPDB (http://EuRBPDB.syshospital.org:8081). We identified a total of 311,571 RBPs with RBDs and 3,639 non-canonical RBPs without known RBDs. EuRBPDB provides detailed annotations for each RBP, including basic information and functional annotation. Moreover, we systematically investigated RBPs in the context of cancer biology based on published literatures and large-scale omics data. To facilitate the exploration of the clinical relevance of RBPs, we additionally designed a cancer web interface to systematically and interactively display the biological features of RBPs in various types of cancers. EuRBPDB has a user-friendly web interface with browse and search functions, as well as data downloading function. We expect that EuRBPDB will be a widely-used resource and platform for the RNA biology community.

bioinformatics↗

bioRxiv2023

Comprehensive single cell aging atlas of mammary tissues reveals shared epigenomic and transcriptomic signatures of aging and cancer

Aging is the greatest risk factor for breast cancer; however, how age-related cellular and molecular events impact cancer initiation is unknown. We investigate how aging rewires transcriptomic and epigenomic programs of mouse mammary glands at single cell resolution, yielding a comprehensive resource for aging and cancer biology. Aged epithelial cells exhibit epigenetic and transcriptional changes in metabolic, pro-inflammatory, or cancer-associated genes. Aged stromal cells downregulate fibroblast marker genes and upregulate markers of senescence and cancer-associated fibroblasts. Among immune cells, distinct T cell subsets (Gzmk+, memory CD4+, {gamma}{delta}) and M2-like macrophages expand with age. Spatial transcriptomics reveal co-localization of aged immune and epithelial cells in situ. Lastly, transcriptional signatures of aging mammary cells are found in human breast tumors, suggesting mechanistic links between aging and cancer. Together, these data uncover that epithelial, immune, and stromal cells shift in proportions and cell identity, potentially impacting cell plasticity, aged microenvironment, and neoplasia risk.

genomics↗

bioRxiv2017

Classification Of Gene Signatures For Their Information Value And Functional Redundancy

Large collections of gene signatures play a pivotal role in interpreting results of omics data analysis but suffer from compositional (large overlap) and functional (redundant read-outs) redundancy, and many gene signatures rarely pop-up in statistical tests. Based on pan-cancer data analysis, here we define a restricted set of 962 so called informative signatures and demonstrate that they have more chances to appear highly enriched in cancer biology studies. We show that the majority of informative signatures conserve their weights for the composing genes (eigengenes) from one cancer type to another. We construct InfoSigMap, an interactive online map showing the structure of compositional and functional redundancies between informative signatures and charting the territories of biological functions accessible through transcriptomic studies. InfoSigMap can be used to visualize in one insightful picture the results of comparative omics data analyses and suggests reconsidering existing annotations of certain reference gene set groups.

systems biology↗

bioRxiv2023

Using graph-based model to identify cell specific synthetic lethal effects

Synthetic lethal (SL) pairs are pairs of genes whose simultaneous loss-of-function results in cell death, while a damaging mutation of either gene alone does not affect the cells survival. This makes SL pairs attractive targets for precision cancer therapies, as targeting the unimpaired gene of the SL pair can selectively kill cancer cells that already harbor the impaired gene. Limited by the difficulty of finding true SL pairs, especially on specific cell types, the identification of SL targets still relies on expensive, time-consuming experimental approaches. In this work, we utilized various cell-line specific omics data to design a deep learning model for predicting SL pairs on particular cell-lines. By incorporating multiple types of cell-specific omics data with a self-attention module, we represent gene relationships as graphs. Our approach demonstrates the potential to facilitate the discovery of cell-specific SL targets for cancer therapeutics, providing a tool to unearth mechanisms underlying the origin of SL in cancer biology. Our approach allows for prediction of SL pairs in a cell-specific manner and enhances cancer precision medicine. The code and data of our approach can be found at https://github.com/promethiume/SLwise HighlightsO_LIFew computational methods can systematically predict SL pairs at a cell-specific level, and their performance may not generalize well to clinical scenarios due to the heterogeneity of cancer types. C_LIO_LIThe SLWise utilizes various cell-line specific omics data to design a deep learning model with a graph-based representation and self-attention mechanism. C_LIO_LIThis approach allows for the prediction of SL pairs in a cell-specific manner, providing valuable insights on effectively identifying the cell-type specific SL targets for personalized treatment strategies. C_LI

bioinformatics↗

bioRxiv2025

GOALS: Gene Ontology Analysis with Layered Shells for Enhanced Functional Insight and Visualization

Gene Ontologies (GOs) are standardized descriptions of gene functions in terms of biological processes, molecular functions, and cellular components, capturing their Parent-Child relationships in a structured framework and advancing cancer biological modeling to provide consistent and meaningful insights into functional genomics analysis. The conventional GO hierarchical structure is defined by human curation experts, with levels determined by the shortest path to the root term. However, grouping GOs poses challenges due to the uneven distribution of gene members within GO terms and inconsistencies in the level of detail across terms at the same GO level. In this work, we introduce Gene Ontology Analysis using Layered Shells (GOALS), a novel tool that discretizes GOAs into optimal layers. GOALS creates scalable GO layers while maintaining a balanced number of genes across GOs in each layer. Unlike existing tools, the GOALS framework organizes GO terms using a bottom-up approach based on their co-membership network, discretizing GOs to achieve an exponential fit with GOs gene member size. Meanwhile, GOALS reveals clusters or supersets reflecting biological relevance by unsupervised clustering of GOs latent projections. In a case study on mouse natural killer (NK) cell development, GOALS identified distinct GO functional clusters with multi-GO layers to reveal multiple levels of detail from specific to abstract contexts to maximize signal discovery and uncover those signals associations with trajectory divergence. More importantly, GOALS enhances enrichment analysis by introducing additional GO stratification and latent GO map that enables more accurate classification of functional differences. GOALS offers a robust and innovative framework for exploring disordered GO clusters, mining GO activities, and analyzing potential GO-GO interplays. By addressing critical challenges in functional genomics, GOALS provides a powerful tool for advancing our understanding of cell heterogeneity and potentially uncovering actionable insights for therapeutic development.

bioinformatics↗

bioRxiv2025

Simultaneous detection and quantification of adenine nucleotides in mammalian cultured cells by HPLC

Adenine nucleotides, including ATP, ADP, ADP-ribose, AMP, NAD+ and NADH, play central roles in cellular homeostasis and involved in multiple metabolic and signaling pathways. Owing to their broad functional relevance in cell biology, the accurate quantification of these metabolites is essential for diverse research areas such as bioenergetics, cell signaling and cancer biology. Several analytical methods have been described for the measurement of adenine nucleotides, ranging from enzymatic assays to mass spectrometry-based approaches. In this study, we developed a reverse-phase high-performance liquid chromatography (RP-HPLC) method with UV-Vis detection that enables the simultaneous quantification of ATP, ADP, ADP-ribose, AMP, NAD+ and NADH. This method is simple, sensitive within the physiological concentration range of all analytes, and capable of detecting biologically relevant changes in ATP, AMP and NAD+ levels induced by pharmacological treatments. Therefore, it presents an accessible and reliable alternative for quantification of these nucleotides in biological samples.

biochemistry↗

bioRxiv2024

Information theoretic inference of magnitude and direction of gene flow in metapopulation networks using nyemtaay, with potential for applications in metastasizing cancer clonal cell origin analysis

BackgroundWe introduce nyemtaay, a Python package for the calculation of classical population genetic statistics and inference of gene flow network connections and directionality in metapopulation networks using information theory. This genetic information flow network inference approach provided here is the only existing implementation of [1], and is applicable not only to ecological and evolutionary organism and landscape scale studies, but also has potential applications in, for example, cancer biology for analyzing clonal cell origins in metastasizing tumors. ResultsWe demonstrate this potential through simulations and an analysis of metastasizing cancer cell lineages, showcasing its ability to identify the tissue site of origin in cancer networks. This work highlights the importance of considering demographic history and founder effects in interpreting gene flow directionality, and the benefits of this understanding in allowing application of this approach to gene flow network modeling to reach a broader range of domains. Conclusionsnyemtaay is available under the MIT license from its public repository (https://github.com/aortizsax/nyemtaay), and can be installed locally using the Python package manager pip.

bioinformatics↗

bioRxiv2026

3D Droplet-Based Bioprinting of Customized In Vitro Head and Neck Cancer Tumor Microenvironment Models

In vitro models are increasingly critical for interrogating cancer biology and therapeutic response, however, accurately recapitulating the tumor microenvironment (TME) remains a persistent challenge, particularly in head and neck cancers (HNC) characterized by complex cell-matrix interactions and heterogeneity. Current models often lack independent tunability of biochemical and biophysical cues, limiting systematic investigation of microenvironmental cues in a high-throughput format. Here, we establish a 3D droplet-based bioprinting platform for the fabrication of customizable in vitro TME models using poly(ethylene glycol) (PEG) hydrogels. Human HNC cell lines (FaDu and 2A3) with differing HPV statuses were bioprinted into PEG matrices spanning physiologically relevant stiffnesses (0.7-4.8 kPa) and compositions, including non-functionalized PEG and peptide-functionalized PEG (PEGfnc: RGD, YIGSR, CNYYSNS) and cultured for 7 days. Cluster growth, cell viability, and cluster morphology were assessed across multiple time points, matrix compositions, and matrix stiffnesses. Proliferation and endpoint phenotype expression were visualized using confocal microscopy through immunofluorescence. Results indicated enhanced cell viability in PEGfnc matrices, compared to non-functionalized matrices, while effect of matrix stiffness was less prominent. Median cluster size reached 40-50 m by day 7, and linear mixed-effects modeling identified how changes in cluster surface area, volume, and tumoroid complexity varied with cell type, matrix, and stiffness. By decoupling and systematically varying key TME parameters, this approach provides a robust and scalable framework for dissecting tumor-matrix interactions and advancing physiologically relevant in vitro models for cancer research and therapeutic screening.

bioengineering↗

bioRxiv2025

Cross-Species Insights from ART-D to Uncover Evolutionarily Conserved Oncogenic Mechanisms

Cancer arises from oncogenic clones, yet the dynamic mechanisms governing their stepwise evolution toward malignancy remain incompletely understood. Here, we establish the Atlas of Ras-driven Tumors in Drosophila (ART-D), a systematic, cross-species platform that dissects the molecular and phenotypic trajectories of tumorigenesis through ten genetically defined RasV12-driven models. By integrating longitudinal phenotypic profiling, we define three conserved stages of tumor development--initiation, promotion, and progression--distinguished by dynamic changes in tumor burden and tumor-induced cachexia. Transcriptomic dynamics reveal stage-specific signaling rewiring: early tumorigenesis is marked by co-activation of JAK/STAT, NF-{kappa}B/Toll, and MAPK pathways, whereas malignant progression is driven by Notch hyperactivation and Hippo pathway inactivation. Through integrative multi-omics and machine learning, we uncover an evolutionarily conserved pathogenic network involving coordinated JNK, NF-{kappa}B/Toll, Notch, and Hippo signaling that is functionally validated across species. ART-D provides a transformative resource that bridges Drosophila genetics and human cancer biology, offering a framework for decoding conserved oncogenic principles and enabling precision targeting of stage-specific vulnerabilities in RAS-driven cancers.

genetics↗

bioRxiv2020

Single-cell nucleic acid profiling in droplets (SNAPD) enables high-throughput analysis of heterogeneous cell populations

Experimental methods that capture the individual properties of single cells are revealing the key role of cell-to-cell variability in countless biological processes. These single-cell methods are becoming increasingly important across the life sciences in fields such as immunology, regenerative medicine, and cancer biology. Existing single-cell analysis methods are often limited by their low analysis throughput, their inability to profile high-dimensional phenotypes, and complicated experimental workflows with slow turnaround times. In this work, we present Single-cell Nucleic Acid Profiling in Droplets (SNAPD) to analyze the transcriptional states of hundreds of thousands of single mammalian cells. Individual cells are encapsulated in aqueous droplets on a microfluidic chip and the content of each cell is profiled by amplifying a targeted panel of transcriptional markers. Molecular logic circuits then integrate this multi-dimensional information to categorize cells based on their transcriptional profile and produce a detectable fluorescence output. SNAPD analyzes over 100,000 cells per hour and can be used to quantify distinct cell types within populations, detect rare cells at frequencies down to 0.1%, and enrich specific cell types using microfluidic sorting. SNAPD provides a simple, rapid, low cost, and scalable approach to study complex phenotypes in heterogeneous cell populations.

bioengineering↗

bioRxiv2018

Comprehensive transcriptomic analysis of cell lines as models of primary tumor samples across 22 tumor types

Cancer cell lines are commonly used as models for cancer biology. While they are limited in their ability to capture complex interactions between tumors and their surrounding environment, they are a cornerstone of cancer research and many important findings have been discovered utilizing cell line models. Not all cell lines are appropriate models of primary tumors, however, which may contribute to the difficulty in translating in vitro findings to patients. Previous studies have leveraged public datasets to evaluate cell lines as models of primary tumors, but they have been limited in scope to specific tumor types and typically ignore the presence of tumor infiltrating cells in the primary tumor samples. We present here a comprehensive pan-cancer analysis utilizing approximately 9,000 transcriptomic profiles from The Cancer Genome Atlas and the Cancer Cell Line Encyclopedia to evaluate cell lines as models of primary tumors across 22 different tumor types. After adjusting for tumor purity in the primary tumor samples, we performed correlation analysis and differential gene expression analysis between the primary tumor samples and cell lines. We found that cell-cycle pathways are consistently upregulated in cell lines, while no pathways are consistently upregulated across the primary tumor samples. In a case study, we compared colorectal cancer cell lines with primary tumor samples across the colorectal subtypes and identified three colorectal cell lines that were derived from fibroblasts rather than tumor epithelial cells. Lastly, we propose a new set of cell lines panel, the TCGA-110, which contains the most representative cell lines from 22 different tumor types as a more comprehensive and informative alternative to the NCI-60 panel. Our analysis of the other tumor types are available in our web app (http://comphealth.ucsf.edu/TCGA110) as a resource to the cancer research community, and we hope it will allow researchers to select more appropriate cell line models and increase the translatability of in vitro findings.

bioinformatics↗

bioRxiv2023

Multiplexed tumor profiling with generative AI accelerates histopathology workflows and improves clinical predictions

Understanding the spatial heterogeneity of tumors and its links to disease is a cornerstone of cancer biology. Emerging spatial technologies offer unprecedented capabilities towards this goal, but several limitations hinder their clinical adoption. To date, histopathology workflows still heavily depend on hematoxylin & eosin (H&E) and serial immunohistochemistry (IHC) staining, a cumbersome and tissue-exhaustive process that yields unaligned tissue images. We propose the VirtualMultiplexer, a generative AI toolkit that translates real H&E images to matching IHC images for several markers based on contrastive learning. The VirtualMultiplexer learns from unpaired H&E and IHC images and introduces a novel multi-scale loss to ensure consistent and biologically reliable stainings. The virtually multiplexed images enabled training a Graph Transformer that simultaneously learns from the joint spatial distribution of several markers to predict clinically relevant endpoints. Our results indicate that the VirtualMultiplexer achieves rapid, robust and precise generation of virtually multiplexed imaging datasets of high staining quality that are indistinguishable from the real ones. We successfully employed transfer learning to generate realistic virtual stainings across tissue scales, patient cohorts, and cancer types with no need for model fine-tuning. Crucially, the generated images are not only realistic but also clinically relevant, as they greatly improved the prediction of different clinical endpoints across patient cohorts and cancer types, speeding up histopathology workflows and accelerating spatial biology.

bioinformatics↗

bioRxiv2025

Decoding Cell Cycle Phase Variations in Cancer Hallmarks Across Breast Cancer Subgroups

Breast tumors have traditionally been categorized into four widely recognized subtypes. Recent single-cell advancements have, however, shown that individual tumors can contain more than one subtype, exhibiting their intrinsic heterogeneity. Relatedly, whilst the crucial role of cell cycle regulation in cancer development is widely acknowledged, the specific impact of cell cycle phases on oncogenic signaling and processes in each of these subtypes remains unresolved. Here, we examine the differences in transcriptomic cancer hallmarks across and within cell cycle phases and breast cancer subtypes in two large single-cell cohorts. We show that taking cell cycle phases into account identifies a broader pool of relevant cancer hallmark pathways while preserving the hallmarks found when cell cycle phases are ignored. Moreover, we identify FDA approved drugs that can target diverse hallmarks within specific phase-subtype cell combinations. This study highlights the critical role of cell cycle phases in advancing our understanding of breast cancer biology. The distinct biological mechanisms observed in each phase-subtype group suggest that targeting these hallmarks may provide a more precise and effective therapeutic strategy for breast cancer treatment.

bioinformatics↗

bioRxiv2025

SURFR: genome-free discovery of cancer-unique small RNAs from cohort sequencing data

Small RNAs play numerous roles in cancer biology and function as biomarkers and therapeutic targets. A recently discovered class of small RNAs - the orphan non-coding RNAs (oncRNAs) - are uniquely present in cancer and absent from healthy tissues. Here, we developed a reference-free computational approach that is independent from the human genome and applied it on cohort sequencing data from five cancer types to discover 1211 novel oncRNAs. Most of these new sequences do not map perfectly to the human genome, and hundreds correlate with patient outcomes. Our new oncRNAs include the first bona fide cancer-unique microRNA; oncRNAs predictive of patient survival and a small RNA that is expressed from an evolutionarily ancestral part of the genome that has been lost in most human individuals. In summary, we provide our reference-free computational approach and more than thousand new cancer-unique RNAs as resources to the community.

genomics↗