Biology subjects

Monti, S.

Publications and source records attributed to Monti, S..

At least 37 records · Page 2Linked to original sources

bioRxiv2024

Insulin Resistance Increases TNBC Aggressiveness and Brain Metastasis via Adipocyte-derived Exosomes

Patients with triple negative breast cancer (TNBC) and comorbid Type 2 Diabetes (T2D), characterized by insulin resistance of adipose tissue, have higher risk of metastasis and shorter survival. Adipocytes are the main non-malignant cells of the breast tumor microenvironment (TME). However, adipocyte metabolism is usually ignored in oncology and mechanisms that couple T2D to TNBC outcomes are poorly understood. Here we hypothesized that exosomes, small vesicles secreted by TME breast adipocytes, drive epithelial-to-mesenchymal transition (EMT) and metastasis in TNBC via miRNAs. Exosomes were purified from conditioned media of 3T3-L1 mature adipocytes, either insulin-sensitive (IS) or insulin-resistant (IR). Murine 4T1 cells, a TNBC model, were treated with exosomes in vitro (72h). EMT, proliferation and angiogenesis were elevated in IR vs. control and IS. Brain metastases showed more mesenchymal morphology and EMT enrichment in the IR group. MiR-145a-3p is highly differentially expressed between IS and IR, and potentially regulates metastasis. SignificanceIR adipocyte exosomes modify TME, increase EMT and promote metastasis to distant organs, likely through miRNA pathways. We suggest metabolic diseases such as T2D reshape the TME, promoting metastasis and decreasing survival. Therefore, TNBC patients with T2D should be closely monitored for metastasis, with metabolic medications considered.

cancer biology↗

bioRxiv2024

Cross-platform proteomics signatures of extreme old age

In previous work we used a Somalogic platform targeting approximately 5000 proteins to generate a serum protein signature of centenarians that we validated in independent studies that used the same technology. We set here to validate and possibly expand the results by profiling the serum proteome of a subset of individuals included in the original study using liquid chromatography tandem mass spectrometry (LC-MS/MS). Following pre-processing, the LC-MS/MS data provided quantification of 398 proteins, with only 266 proteins shared by both platforms. At 1% FDR statistical significance threshold, the analysis of LC-MS/MS data detected 44 proteins associated with extreme old age, including 23 of the original analysis. To identify proteins for which associations between expression and extreme-old age were conserved across platforms, we performed inter-study conservation testing of the 266 proteins quantified by both platforms using a method that accounts for the correlation between the results. From these tests, a total of 80 proteins reached 5% FDR statistical significance, and 26 of these proteins had concordant pattern of gene expression in whole blood. This signature of 80 proteins points to blood coagulation, IGF signaling, extracellular matrix (ECM) organization, and complement cascade as important pathways whose protein level changes provide evidence for age-related adjustments that distinguish centenarians from younger individuals.

bioinformatics↗

bioRxiv2024

Learning Gaussian Graphical Models from Correlated Data

Gaussian Graphical Models (GGM) have been widely used in biomedical research to explore complex relationships between many variables. There are well established procedures to build GGMs from a sample of independent and identical distributed observations. However, many studies include clustered and longitudinal data that result in correlated observations and ignoring this correlation among observations can lead to inflated Type I error. In this paper, we propose a Bootstrap algorithm to infer GGM from correlated data. We use extensive simulations of correlated data from family-based studies to show that the Bootstrap method does not inflate the Type I error while retaining statistical power compared to alternative solutions. We apply our method to learn the GGM that represents complex relations between 47 Polygenic Risk Scores generated using genome-wide genotype data from a family-based study known as the Long Life Family Study. By comparing it to the conventional methods that ignore within-cluster correlation, we show that our method controls the Type I error well in this real example.

bioinformatics↗

bioRxiv2024

yQTL Pipeline: a structured computational workflow for large scale quantitative trait loci discovery and downstream visualization

1Quantitative trait loci (QTL) denote regions of DNA whose variation is associated with variations in quantitative traits. QTL discovery is a powerful approach to understand how changes in molecular and clinical phenotypes may be related to DNA sequence changes. However, QTL discovery analysis encompasses multiple analytical steps and the processing of multiple input files, which can be laborious, error prone, and hard to reproduce if performed manually. In order to facilitate and automate large-scale QTL analysis, we developed the yQTL Pipeline, where the y indicates the dependent quantitative variable being modeled. Prior to genome-wide association test, the pipeline supports the calculation or the direct input of pre-defined genome-wide principal components and genetic relationship matrix when applicable. User-specified covariates can also be provided. Depending on whether familial relatedness exists among the subjects, genome-wide association tests will be performed using either a linear mixed-effect model or a linear model. Using the workflow management tool Nextflow, the pipeline parallelizes the analysis steps to optimize run-time and ensure results reproducibility. In addition, a user-friendly R Shiny App is developed to facilitate result visualization. Upon uploading the result file, it can generate Manhattan plots of user-selected phenotype traits and trait-QTL connection networks based on user-specified p-value thresholds. We applied the yQTL Pipeline to analyze metabolomics profiles of blood serum from the New England Centenarians Study (NECS) participants. A total of 9.1M SNPs and 1,052 metabolites across 194 participants were analyzed. Using a p-value cutoff 5e-8, we found 14,983 mQTLs cumulatively associated with 312 metabolites. The built-in parallelization of our pipeline reduced the run time from [~]90 min to [~]26 min. Visualization using the R Shiny App revealed multiple mQTLs shared across multiple metabolites. The yQTL Pipeline is available with documentation on GitHub at https://github.com/montilab/yQTL-Pipeline.

bioinformatics↗

bioRxiv2023

Total RNA sequencing reveals gene expression and microbial alterations shared by oral pre-malignant lesions and cancer.

Head and neck cancers are a complex malignancy comprising multiple anatomical sites, with cancer of the oral cavity ranking among the deadliest and most disfiguring cancers globally. Oral cancer (OC) constitutes a subset of head and neck cancer cases, presenting primarily as tobacco-and alcohol-associated oral squamous cell carcinoma (OSCC), with a 5-year survival rate of [~]65%, partly due to the lack of early detection and effective treatments. OSCC arises from premalignant lesions (PMLs) in the oral cavity through a multi-step series of clinical and histopathological stages, including varying degrees of epithelial dysplasia. To gain insights into the molecular mechanisms associated with the progression of PMLs to OSCC, we profiled the whole transcriptome of 66 human PMLs comprising leukoplakia with dysplasia and hyperkeratosis non-reactive (HkNR) pathologies, alongside healthy controls and OSCC. Our data revealed that PMLs were enriched in gene signatures associated with cellular plasticity, such as partial EMT (p-EMT) phenotypes, and with immune response. Integrated analyses of the host transcriptome and microbiome further highlighted a significant association between differential microbial abundance and PML pathway activity, suggesting a contribution of the oral microbiome towards PML evolution to OSCC. Collectively, this study reveals molecular processes associated with PML progression that may help early diagnosis and disease interception at an early stage. AUTHOR SUMMARYPatients harboring oral premalignant lesions (PMLs) have an increased risk of developing oral squamous cell carcinoma (OSCC), but the underlying mechanisms driving transformation of PMLs to OSCC remain poorly understood. In this study, Khan et al., analyzed a newly generated dataset of gene expression and microbial profiles of oral tissues from patients diagnosed with PMLs from differing histopathological groups, including hyperkeratosis not reactive (HkNR) and dysplasia, comparing these profiles with OSCC and normal oral mucosa. Significant similarities between PMLs and OSCC were observed, with PMLs manifesting several cancer hallmarks, including oncogenic and immune pathways. The study also demonstrates associations between the abundance of multiple microbial species and PML groups, suggesting a potential contribution of the oral microbiome to the early stages of OSCC development. The study offers insights into the nature of the molecular, cellular and microbial heterogeneity of oral PMLs and suggests that molecular and clinical refinement of PMLs may provide opportunities for early disease detection and interception. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=94 SRC="FIGDIR/small/534064v1_ufig1.gif" ALT="Figure 1"> View larger version (27K): org.highwire.dtl.DTLVardef@86cf7org.highwire.dtl.DTLVardef@129e6d1org.highwire.dtl.DTLVardef@1f0ae31org.highwire.dtl.DTLVardef@3c4d67_HPS_FORMAT_FIGEXP M_FIG C_FIG

genomics↗

bioRxiv2023

Bayesian Differential Analysis of Cell Type Proportions

The analysis of cell type proportions in a biological sample should account for the compositional nature of the data but most analyses ignore this characteristic with the risk of producing misleading conclusions. The recent method scCODA appropriately incorporates these constraints by using a Bayesian Multinomial-Dirichlet model that requires a reference cell type to normalize the distribution of all cell types. However, a reference cell type that is stable across biological conditions may not always be available. Here, we present an approach that uses a Bayesian multinomial regression for the analysis of single cell distribution data without the need for a reference cell type. We show an implementation example using the rjags package within the R software.

bioinformatics↗

bioRxiv2022

Multi-modal profiling of peripheral blood cells across the human lifespan reveals distinct immune cell signatures of aging and longevity

Age-related changes in immune cell composition and functionality are associated with multimorbidity and mortality. However, many centenarians delay the onset of aging-related disease suggesting the presence of elite immunity that remains highly functional at extreme old age. To identify immune-specific patterns of aging and extreme human longevity, we analyzed novel single cell profiles from the peripheral blood mononuclear cells (PBMCs) of 7 centenarians (mean age 106) and publicly available single cell RNA-sequencing (scRNA-seq) datasets that included an additional 7 centenarians as well as 52 people at younger ages (20-89 years). The analysis confirmed known shifts in the ratio of lymphocytes to myeloid cells, and noncytotoxic to cytotoxic cell distributions with aging, but also identified significant shifts from CD4+ T cell to B cell populations in centenarians suggesting a history of exposure to natural and environmental immunogens. Our transcriptional analysis identified cell type signatures specific to exceptional longevity that included genes with age-related changes (e.g., increased expression of STK17A, a gene known to be involved in DNA damage response) as well as genes expressed uniquely in centenarians PBMCs (e.g., S100A4, part of the S100 protein family studied in age-related disease and connected to longevity and metabolic regulation). Collectively, these data suggest that centenarians harbor unique, highly functional immune systems that have successfully adapted to a history of insults allowing for the achievement of exceptional longevity.

immunology↗

bioRxiv2021

Structure Learning for Hierarchical Regulatory Networks

Network analysis offers a powerful technique to model the relationships between genes within biological regulatory networks. Inference of biological network structures is often performed on high-dimensional data, yet is hindered by the limited sample size of high throughput "omics" data typically available. To overcome this challenge, we exploit known organizing principles of biological networks that are sparse, modular, and likely share a large portion of their underlying architecture. We present SHINE - Structure Learning for Hierarchical Networks - a framework for defining data-driven structural constraints and incorporating a shared learning paradigm for efficiently learning multiple networks from high-dimensional data. We show through simulations SHINE improves performance when relatively few samples are available and multiple networks are desired, by reducing the complexity of the graphical search space and by taking advantage of shared structural information. We evaluated SHINE on TCGA Pan-Cancer data and found learned tumor-specific networks exhibit expected graph properties of real biological networks, recapture previously validated interactions, and recapitulate findings in literature. Application of SHINE to the analysis of subtype-specific breast cancer networks identified key genes and biological processes for tumor maintenance and survival as well as potential therapeutic targets for modulating known breast cancer disease genes.

bioinformatics↗

bioRxiv2020

Multi-resolution characterization of molecular taxonomies in bulk and single-cell transcriptomics data

As high-throughput genomics assays become more efficient and cost effective, their utilization has become standard in large-scale biomedical projects. These studies are often explorative, in that relationships between samples are not explicitly defined a priori, but rather emerge from data-driven discovery and annotation of molecular subtypes, thereby informing hypotheses and independent evaluation. Here, we present K2Taxonomer, a novel unsupervised recursive partitioning algorithm and associated R package that utilize ensemble learning to identify robust subgroups in a "taxonomy-like" structure (https://github.com/montilab/K2Taxonomer). K2Taxonomer was devised to accommodate different data paradigms, and is suitable for the analysis of both bulk and single-cell transcriptomics data. For each of these data types, we demonstrate the power of K2Taxonomer to discover known relationships in both simulated and human tissue data. We conclude with a practical application on breast cancer tumor infiltrating lymphocyte (TIL) single-cell profiles, in which we identified co-expression of translational machinery genes as a dominant transcriptional program shared by T cells subtypes, associated with better prognosis in breast cancer tissue bulk expression data.

bioinformatics↗

bioRxiv2020

animalcules: Interactive Microbiome Analytics 1 and Visualization in R

BackgroundMicrobial communities that live in and on the human body play a vital role in health and disease. Recent advances in sequencing technologies have enabled the study of microbial communities at unprecedented resolution. However, these advances in data generation have presented novel challenges to researchers attempting to analyze and visualize these data. ResultsTo address some of these challenges, we have developed animalcules, an easy-to-use interactive microbiome analysis toolkit for 16S rRNA sequencing data, shotgun DNA metagenomics data, and RNA-based metatranscriptomics profiling data. This toolkit combines novel and existing analytics, visualization methods, and machine learning models. For example, traditional microbiome analyses such as alpha/beta diversity and differential abundance analysis are enhanced in the toolkit, while new methods such as biomarker identification are introduced. Powerful interactive and dynamic figures generated by animalcules enable users to understand their data and discover new insights. animalcules can be used as a standalone command-line R package or users can explore their data with the accompanying interactive R Shiny interface. ConclusionsWe present animalcules, an R package for interactive microbiome analysis through either an interactive interface facilitated by R Shiny or various command-line functions. It is the first microbiome analysis toolkit that supports the analysis of all 16S rRNA, DNA-based shotgun metagenomics, and RNA-sequencing based metatranscriptomics datasets. animalcules can be freely downloaded from GitHub at https://github.com/compbiomed/animalcules or installed through Bioconductor at https://www.bioconductor.org/packages/release/bioc/html/animalcules.html.

microbiology↗

bioRxiv2019

hypeR: An R Package for Geneset Enrichment Workflows

SummaryGeneset enrichment is a popular method for annotating high-throughput sequencing data. Existing tools fall short in providing the flexibility to tackle the varied challenges researchers face in such analyses, particularly when analyzing many signatures across multiple experiments. We present a comprehensive R package for geneset enrichment workflows that offers multiple enrichment, visualization, and sharing methods in addition to novel features such as hierarchical geneset analysis and built-in markdown reporting. hypeR is a one-stop solution to performing geneset enrichment for a wide audience and range of use cases.\n\nAvailability and implementationThe most recent version of the package is available at https://github.com/montilab/hypeR.\n\nSupplementary informationComprehensive documentation and tutorials, are available at https://montilab.github.io/hypeR-docs.

bioinformatics↗

bioRxiv2019

Yap suppresses T cell function and infiltration in the tumor microenvironment

A major challenge for cancer immunotherapy is sustaining T cell activation and recruitment in immunosuppressive solid tumors. Here we report that Yap levels are sharply induced upon activation of CD4+ and CD8+ T cells and that Yap functions as an immunosuppressive factor and inhibitor of effector differentiation. Loss of Yap in T cells results in enhanced T cell activation, differentiation and function, which translates in vivo to an improved ability for T cells to infiltrate and repress tumors. Gene expression analyses of tumor-infiltrating T cells following Yap deletion implicates Yap as a mediator of global T cell responses in the tumor microenvironment and as a key negative regulator of T cell tumor infiltration and patient survival in diverse human cancers. Collectively, our results indicate that Yap plays critical roles in T cell biology, and suggest that inhibiting Yap activity improves T cell responses in cancer.

cancer biology↗

bioRxiv2019

iEDGE: integration of Epi-DNA and Gene Expression and applications to the discovery of somatic copy number-associated drivers in cancer

The emergence of large-scale multi-omics data warrants method development for data integration. Genomic studies from cancer patients have identified epigenetic and genetic regulators - such as methylation marks, somatic mutations, and somatic copy number alterations (SCNAs), among others - as predictive features of cancer outcome. However, identification of \"driver genes\" associated with a given alteration remains a challenge. To this end, we developed a computational tool, iEDGE, to model cis and trans effects of (epi-)DNA alterations and identify potential cis driver genes, where cis and trans genes denote those genes falling within and outside the genomic boundaries of a given (epi-)genetic alteration, respectively.\n\nFirst, iEDGE identifies the cis and trans genes associated with the presence/absence of a particular epi-DNA alteration across samples. Tests of statistical mediation are then performed to determine the cis genes predictive of the trans gene expression. Finally, cis and trans effects are annotated by pathway enrichment analysis to gain insights into the underlying regulatory networks.\n\nWe used iEDGE to perform integrative analysis of SCNAs and gene expression data from breast cancer and 18 additional cancer types included in The Cancer Genome Atlas (TCGA). Notably, cis gene drivers identified by iEDGE were found to be significantly enriched for known driver genes from multiple compendia of validated oncogenes and tumor suppressors, suggesting that the remainder are of equal importance. Furthermore, predicted drivers were enriched for functionally relevant cancer genes with amplification-driven dependencies, which are of potential prognostic and therapeutic value. All the analyses results are accessible at https://montilab.bu.edu/iEDGE.

bioinformatics↗

bioRxiv2019

A Data-Driven Transcriptional Taxonomy of Adipogenic Chemicals to Identify Emerging Metabolic Health Threats

BackgroundChemicals in disparate structural classes activate specific subsets of PPAR{gamma}s transcriptional programs to generate adipocytes with distinct phenotypes. ObjectivesOur objectives were to 1) establish a novel classification method to predict PPAR{gamma} ligands and modifying chemicals, and 2) create a taxonomy to group chemicals based on their effects on PPAR{gamma}s transcriptome and downstream metabolic functions. We tested the hypothesis that environmental adipogens highly ranked by the taxonomy, but segregated from therapeutic PPAR{gamma} ligands, would induce white but not brite adipogenesis. Methods3T3-L1 cells were differentiated in the presence of 76 chemicals (negative controls, nuclear receptor ligands known to influence adipocyte biology, potential environmental PPAR{gamma} ligands). Differentiation was assessed by measuring lipid accumulation. mRNA expression was determined by RNA-Seq and validated by RT-qPCR. A novel classification model was developed using an amended random forest procedure. A subset of environmental contaminants identified as strong PPAR{gamma} agonists were analyzed by their effects on lipid handling, mitochondrial biogenesis and cellular respiration in 3T3-L1 cells and human preadipocytes. ResultsWe used lipid accumulation and RNA sequencing data to develop a classification system that 1) identified PPAR{gamma} agonists, and 2) sorted chemicals into likely white or brite adipogens. Expression of Cidec was the most efficacious indicator of strong PPAR{gamma} activation. Two known environmental PPAR{gamma} ligands, tetrabromobisphenol A and triphenyl phosphate, which sorted distinctly from therapeutic ligands, induced white adipocyte genes but failed to induce Pgc1a and Ucp1, and induced fatty acid uptake but not mitochondrial biogenesis in 3T3-L1 cells. Moreover, two chemicals identified as highly ranked PPAR{gamma} agonists, tonalide and quinoxyfen, induced white adipogenesis without the concomitant health-promoting characteristics of brite adipocytes in mouse and human preadipocytes. DiscussionA novel classification procedure accurately identified environmental chemicals as PPAR{gamma} ligands distinct from known PPAR{gamma}-activating therapeutics. The computational and experimental framework has general applicability to the classification of as-yet uncharacterized chemicals.

pharmacology and toxicology↗

bioRxiv2018

Pipeliner: A Nextflow-based framework for the definition of sequencing data processing pipelines

The advent of high-throughput sequencing technologies has led to the need for flexible and user-friendly data pre-processing platforms. The Pipeliner framework provides an out-of-the-box solution for processing various types of sequencing data. It combines the Nextflow scripting language and Anaconda package manager to generate modular computational workflows. We have used Pipeliner to create several pipelines for sequencing data processing including bulk RNA-seq, single-cell RNA-seq (scRNA-seq), as well as Digital Gene Expression (DGE) data. This report highlights the design methodology behind Pipeliner which enables the development of highly flexible and reproducible pipelines that are easy to extend and maintain on multiple computing environments. We also provide a quick start user guide demonstrating how to setup and execute available pipelines with toy datasets.

bioinformatics↗

bioRxiv2018

Assessment of a highly multiplexed RNA sequencing platform and comparison to existing high-throughput gene expression profiling techniques

The need to reduce per sample cost of RNA-seq profiling for scalable data generation has led to the emergence of highly multiplexed RNA-seq. These technologies utilize barcoding of cDNA sequences in order to combine samples into single sequencing lane to be separated during data processing. In this study, we report the performance of one such technique denoted as sparse full length sequencing (SFL), a ribosomal RNA depletion-based RNA sequencing approach that allows for the simultaneous sequencing of 96 samples and higher. We offer comparisons to well established single-sample techniques, including: full coverage Poly-A capture RNA-seq and microarray, as well as another low-cost highly multiplexed technique known as 3 digital gene expression (3DGE). Data was generated for a set of exposure experiments on immortalized human lung epithelial (AALE) cells in a two-by-two study design, in which samples received both genetic and chemical perturbations of known oncogenes/tumor suppressors and lung carcinogens. SFL demonstrated improved performance over 3DGE in terms of coverage, power to detect differential gene expression, and biological recapitulation of patterns of differential gene expression from in vivo lung cancer mutation signatures.

bioinformatics↗

bioRxiv2018

Tributyltin induces a transcriptional response without a brite adipocyte signature in adipocyte models

Tributyltin (TBT), a peroxisome proliferator-activated receptor {gamma} (PPAR{gamma})/retinoid X receptor (RXR) ligand and founding member of the environmental obesogen chemical class, induces adipocyte differentiation and suppresses bone formation. A growing number of environmental PPAR{gamma} ligands are being identified. However, the potential for environmental PPAR{gamma} ligands to induce adverse metabolic effects has been questioned because PPAR{gamma} is a therapeutic target in treatment of type II diabetes. We evaluated the molecular consequences of TBT exposure during bone marrow multipotent mesenchymal stromal cell (BM-MSC) differentiation in comparison to rosiglitazone, a therapeutic PPAR{gamma} ligand, and LG100268, a synthetic RXR ligand. Mouse primary BM-MSCs (female, C57BL/6J) undergoing bone differentiation were exposed to maximally efficacious and human relevant concentrations of rosiglitazone (100 nM), LG100268 (100 nM) or TBT (80 nM) for 4 days. Gene expression was assessed using microarrays, and in silico functional annotation was performed using pathway enrichment analysis approaches. Pathways related to osteogenesis were downregulated by all three ligands, while pathways related to adipogenesis were upregulated by rosiglitazone and TBT. However, pathways related to mitochondrial biogenesis and brown-in-white (brite) adipocyte differentiation were more significantly upregulated in rosiglitazone-treated than TBT-treated cells. The lack of induction of genes involved in adipocyte energy dissipation by TBT was confirmed by an independent gene expression analysis in BM-MSCs undergoing adipocyte differentiation and by analysis of a publically available 3T3 L1 data set. Furthermore, rosiglitazone, but not TBT, induced mitochondrial biogenesis. This study is the first to show that an environmental PPAR{gamma} ligand has a limited capacity to induce health promoting activities of PPAR{gamma}.

pharmacology and toxicology↗

bioRxiv2018

The Carcinogenome Project: In-vitro Gene Expression Profiling of Chemical Perturbations to Predict Long-Term Carcinogenicity

Background: Most chemicals in commerce have not been evaluated for their carcinogenic potential. The current de-facto gold-standard approach to carcinogen testing adopts the two-year rodent bioassay, a time consuming and costly procedure. Alternative approaches, such as high-throughput in-vitro assays, show promise in addressing the limitations in carcinogen screening.\n\nObjectives: We developed a screening process for predicting chemical carcinogenicity and genotoxicity and characterizing modes of actions (MoAs) using in-vitro gene expression assays.\n\nMethods: We generated a large toxicogenomics resource comprising ~6,000 expression profiles corresponding to 330 chemicals profiled in HepG2 cells at multiple doses and in replicates. Predictive models of carcinogenicity were built using a Random Forest classifier. Differential pathway enrichment analysis was performed to identify pathways associated with carcinogen exposure. Signatures of carcinogenicity and genotoxicity were compared with external data sources including Drugmatrix and the Connectivity Map.\n\nResults: Among profiles with sufficient bioactivity, our classifiers achieved 72.2% AUC for predicting carcinogenicity and 82.3% AUC for predicting genotoxicity. Our analysis showed that chemical bioactivity, as measured by the strength and reproducibility of the transcriptional response, is not significantly associated with long-term carcinogenicity, as evidenced by the many carcinogenic chemicals that did not elicit substantial changes in gene expression at doses up to 40 M. However, sufficiently high transcriptional bioactivity is necessary for a chemical to be used for prediction of carcinogenicity. Pathway enrichment analysis revealed several pathways consistent with literature review of pathways that drive cancer, including DNA damage and DNA repair. These data are available for download via https://clue.io/CRCGN_ABC, and a web portal for interactive query and visualization of the data and results is accessible at https://carcinogenome.org.\n\nConclusions: We demonstrated a short-term in-vitro screening approach using gene expression profiling to predict long-term carcinogenicity and infer MoAs of chemical perturbations.

bioinformatics↗