THU0388 Combining RNA-SEQ and Machine Learning to Classify an Sle-Specific Gene Signature and in Vitro Responses to IFN-I Pathway Inhibition

Abstract Leptomeningeal metastases (LM), a diffuse form of brain metastases is rare and fatal progression of non-small cell lung cancer (NSCLC). In LM, metastatic cancer cells spread and resign on the brain meninges, the cerebrospinal fluid (CSF), cranial and spinal nerves. Rapid disease progression and scarce tissue availability hinder the progress of scientific study of LM and its treatment. To overcome the critical lack of tissue and to determine the genetic profile of NSCLC LM, we have developed methods to extract tumor-associated cell-free RNA from CSF, and isolated and sequenced circulating single cells from CSF. Herein, we used high throughput qPCR to target lung and brain-associated genes and identified NSCLC LM metastases-related RNA. Brain-specific gene signature (GFAP, NRGN, SNCB, ZBTB18) was detected in all CSF sample (control and metastases), whereas lung-specific genes (MUC1, SFTPB, SFTPD, SLC34A2) were detected in CSF of brain metastases patients. Normal, healthy CSF lacks cellular component, but CSF of patients with LM metastases inhabited with very low amount of circulating tumor cells. Single cells from CSF of 4 patients with NSCLC LM metastases were captured with microfluidic chip. Cells (n = 197) were clustered by significantly differential expressed genes demonstrating two distinct populations of white blood and tumor cells. These data identified specific cfRNA and single cell transcriptome profiles compared to normal cells or patients without NSCLC LM metastases, and highlighted metastases-associated carcinoembryonic antigen-related cell adhesion molecule 6 (CEACAM6) as highly expressed in patients with NSCLC LM metastases. CEACAM6 mRNA was detected in CSF of 86% of patients with NSCLC LM but not in the CSF of control patients. In vitro inhibition of CEACAM6 protein lead to decreased invasion in NSCLC cells which was rescued by overexpression of the protein. We have developed sensitive and robust techniques to leverage human CSF to study NSCLC LM.

Download Full-text

A high-throughput RNA-seq approach to profile transcriptional responses

10.1101/018416 ◽

2015 ◽

Cited By ~ 1

Author(s):

Gregory A Moyerbrailean ◽

Gordon O Davis ◽

Chris T Harvey ◽

Donovan Watza ◽

Xiaoquan Wen ◽

...

Keyword(s):

Gene Expression ◽

High Throughput ◽

Deep Sequencing ◽

Transcriptome Profiling ◽

Iterative Refinement ◽

Specific Gene ◽

Rna Seq ◽

Transcriptional Responses ◽

Expression Screening

In recent years, different technologies have been used to measure genome-wide gene expression levels and to study the transcriptome across many types of tissues and in response to in vitro treatments. However, a full understanding of gene regulation in any given cellular and environmental context combination is still missing. This is partly because analyzing tissue/environment-specific gene expression generally implies screening a large number of cellular conditions and samples, without prior knowledge of which conditions are most informative (e.g. some cell types may not respond to certain treatments). To circumvent these challenges, we have established a new two-step high-throughput and cost-effective RNA-seq approach: the first step consists of gene expression screening of a large number of conditions, while the second step focuses on deep sequencing of the most relevant conditions (e.g. largest number of differentially expressed genes). This study design allows for a fast and economical screen in step one, with a more profitable allocation of resources for the deep sequencing of re-pooled libraries in step two. We have applied this approach to study the response to 26 treatments in three lymphoblastoid cell line samples and we show that it is applicable for other high-throughput transcriptome profiling requiring iterative refinement or screening.

Download Full-text

Consensus Machine Learning for Gene Target Selection in Pediatric AML Risk

10.1101/632166 ◽

2019 ◽

Author(s):

Jenny Smith ◽

Sean K. Maden ◽

David Lee ◽

Ronald Buie ◽

Vikas Peddu ◽

...

Keyword(s):

Machine Learning ◽

Random Forest ◽

Expression Patterns ◽

Specific Gene ◽

Molecular Heterogeneity ◽

Learning Approach ◽

Rna Seq ◽

Gene Target ◽

Machine Learning Approach ◽

Pediatric Aml

AbstractAcute myeloid leukemia (AML) is a cancer of hematopoietic systems that poses high population burden, especially among pediatric populations. AML presents with high molecular heterogeneity, complicating patient risk stratification and treatment planning. While molecular and cytogenetic subtypes of AML are well described, significance of subtype-specific gene expression patterns is poorly understood and effective modeling of these patterns with individual algorithms is challenging. Using a novel consensus machine learning approach, we analyzed public RNA-seq and clinical data from pediatric AML patients (N = 137 patients) enrolled in the TARGET project.We used a binary risk classifier (Low vs. Not-Low Risk) to study risk-specific expression patterns in pediatric AML. We applied the following workflow to identify important gene targets from RNA-seq data: (1) Reduce data dimensionality by identification of differentially expressed genes for AML risk (N = 1984 loci); (2) Optimize algorithm hyperparameters for each of 4 algorithm types (lasso, XGBoost, random forest, and SVM); (3) Study ablation test results for penalized methods (lasso and XGBoost); (4) Bootstrap Boruta permutations with a novel consensus importance metric.We observed recurrently selected features across hyperparameter optimizations, ablation tests, and Boruta permutation bootstrap iterations, including HOXA9 and putative cofactors including MEIS1. Consensus feature selection from Boruta bootstraps identified a larger gene set than single penalized algorithm runs (lasso or XGBoost), while also including correlated and predictive genes from ablation tests.We present a consensus machine learning approach to identify gene targets of likely importance for pediatric AML risk. The approach identified a moderately sized set of recurrent important genes from across 4 algorithm types, including genes identified across ablation tests with penalized algorithms (HOXA9 and MEIS1). Our approach mitigates exclusion biases of penalized algorithms (lasso and XGBoost) while obviating arbitrary importance cutoffs for other types (SVM and random forest). This approach is readily generalizable for research of other heterogeneous diseases, single-assay experiments, and high-dimensional data. Resources and code to recreate our findings are available online.

Download Full-text

Human peripheral blood DNAM-1neg NK cells are a terminally differentiated subset with limited effector functions

Blood Advances ◽

10.1182/bloodadvances.2018030676 ◽

2019 ◽

Vol 3 (11) ◽

pp. 1681-1694 ◽

Cited By ~ 8

Author(s):

Kimberley A. Stannard ◽

Sébastien Lemoine ◽

Nigel J. Waterhouse ◽

Frank Vari ◽

Lucienne Chatenoud ◽

...

Keyword(s):

Nk Cells ◽

Peripheral Blood ◽

Nk Cell ◽

Human Peripheral Blood ◽

Cell Subset ◽

B Cell Lymphoma ◽

Gene Signature ◽

Interferon Γ ◽

Specific Gene

Abstract Natural killer (NK) cells are a heterogeneous population of innate lymphocytes whose potent anticancer properties make them ideal candidates for cellular therapeutic application. However, our lack of understanding of the role of NK cell diversity in antitumor responses has hindered advances in this area. In this study, we describe a new CD56dim NK cell subset characterized by the lack of expression of DNAX accessory molecule-1 (DNAM-1). Compared with CD56bright and CD56dimDNAM-1pos NK cell subsets, CD56dimDNAM-1neg NK cells displayed reduced motility, poor proliferation, lower production of interferon-γ, and limited killing capacities. Soluble factors secreted by CD56dimDNAM-1neg NK cells impaired CD56dimDNAM-1pos NK cell–mediated killing, indicating a potential inhibitory role for the CD56dimDNAM-1neg NK cell subset. Transcriptome analysis revealed that CD56dimDNAM-1neg NK cells constitute a new mature NK cell subset with a specific gene signature. Upon in vitro cytokine stimulation, CD56dimDNAM-1neg NK cells were found to differentiate from CD56dimDNAM-1pos NK cells. Finally, we report a dysregulation of NK cell subsets in the blood of patients diagnosed with Hodgkin lymphoma and diffuse large B-cell lymphoma, characterized by decreased CD56dimDNAM-1pos/CD56dimDNAM-1neg NK cell ratios and reduced cytotoxic activity of CD56dimDNAM-1pos NK cells. Altogether, our data offer a better understanding of human peripheral blood NK cell populations and have important clinical implications for the design of NK cell–targeting therapies.

Download Full-text

Integrative Transcriptomic Analysis Reveals Distinctive Molecular Traits and Novel Subtypes of Collecting Duct Carcinoma

Cancers ◽

10.3390/cancers13122903 ◽

2021 ◽

Vol 13 (12) ◽

pp. 2903

Author(s):

Chiara Gargiuli ◽

Pierangela Sepe ◽

Anna Tessari ◽

Tyler Sheetz ◽

Maurizio Colecchia ◽

...

Keyword(s):

Drug Sensitivity ◽

Clear Cell ◽

Collecting Duct ◽

Gene Signature ◽

Specific Gene ◽

Cell Renal Cell Carcinoma ◽

Duct Carcinoma ◽

Collecting Duct Carcinoma ◽

Molecular Features

Collecting duct carcinoma (CDC) is a rare and highly aggressive kidney cancer subtype with poor prognosis and no standard treatments. To date, only a few studies have examined the transcriptomic portrait of CDC. Through integration of multiple datasets, we compared CDC to normal tissue, upper-tract urothelial carcinomas, and other renal cancers, including clear cell, papillary, and chromophobe histologies. Association between CDC gene expression signatures and in vitro drug sensitivity data was evaluated using the Cancer Therapeutic Response Portal, Genomics of Drug Sensitivity in Cancer datasets, and connectivity map. We identified a CDC-specific gene signature that predicted in vitro sensitivity to different targeted agents and was associated to worse outcome in clear cell renal cell carcinoma. We showed that CDC are transcriptionally related to the principal cells of the collecting ducts providing evidence that this tumor originates from this normal kidney cell type. Finally, we proved that CDC is a molecularly heterogeneous disease composed of at least two subtypes distinguished by cell signaling, metabolic and immune-related alterations. Our findings elucidate the molecular features of CDC providing novel biological and clinical insights. The identification of distinct CDC subtypes and their transcriptomic traits provides the rationale for patient stratification and alternative therapeutic approaches.

Download Full-text

Abstract 4501: RNA seq analysis reveals altered immune specific gene signature in African American bladder cancer

10.1158/1538-7445.am2020-4501 ◽

2020 ◽

Author(s):

Karthik reddy kami reddy ◽

Balasubramanyam Karanam ◽

Cristian Coarfa ◽

Yair Lotan ◽

Roni J. Bollag ◽

...

Keyword(s):

Bladder Cancer ◽

African American ◽

Gene Signature ◽

Specific Gene ◽

Rna Seq

Download Full-text

Identification of a Transcriptomic Prognostic Signature by Machine Learning Using a Combination of Small Cohorts of Prostate Cancer

Frontiers in Genetics ◽

10.3389/fgene.2020.550894 ◽

2020 ◽

Vol 11 ◽

Author(s):

Benjamin Vittrant ◽

Mickael Leclercq ◽

Marie-Laure Martin-Magniette ◽

Colin Collins ◽

Alain Bergeron ◽

...

Keyword(s):

Prostate Cancer ◽

Machine Learning ◽

Specific Antigen ◽

Gene Signature ◽

Learning Approach ◽

Gleason Grade ◽

Rna Seq ◽

Number Of Patients ◽

Heterogeneous Datasets ◽

Machine Learning Approach

Determining which treatment to provide to men with prostate cancer (PCa) is a major challenge for clinicians. Currently, the clinical risk-stratification for PCa is based on clinico-pathological variables such as Gleason grade, stage and prostate specific antigen (PSA) levels. But transcriptomic data have the potential to enable the development of more precise approaches to predict evolution of the disease. However, high quality RNA sequencing (RNA-seq) datasets along with clinical data with long follow-up allowing discovery of biochemical recurrence (BCR) biomarkers are small and rare. In this study, we propose a machine learning approach that is robust to batch effect and enables the discovery of highly predictive signatures despite using small datasets. Gene expression data were extracted from three RNA-Seq datasets cumulating a total of 171 PCa patients. Data were re-analyzed using a unique pipeline to ensure uniformity. Using a machine learning approach, a total of 14 classifiers were tested with various parameters to identify the best model and gene signature to predict BCR. Using a random forest model, we have identified a signature composed of only three genes (JUN, HES4, PPDPF) predicting BCR with better accuracy [74.2%, balanced error rate (BER) = 27%] than the clinico-pathological variables (69.2%, BER = 32%) currently in use to predict PCa evolution. This score is in the range of the studies that predicted BCR in single-cohort with a higher number of patients. We showed that it is possible to merge and analyze different small and heterogeneous datasets altogether to obtain a better signature than if they were analyzed individually, thus reducing the need for very large cohorts. This study demonstrates the feasibility to regroup different small datasets in one larger to identify a predictive genomic signature that would benefit PCa patients.

Download Full-text

Epithelial Organotypic Cultures: A Viable Model to Address Mechanisms of Carcinogenesis by Epitheliotropic Viruses

Current Topics in Medicinal Chemistry ◽

10.2174/1568026618666180410125850 ◽

2018 ◽

Vol 18 (4) ◽

pp. 246-255 ◽

Cited By ~ 1

Author(s):

Lara Termini ◽

Enrique Boccardo

Keyword(s):

Three Dimensional ◽

Cell Types ◽

Experimental Models ◽

Biological Research ◽

Specific Gene ◽

Specific Cell ◽

Organotypic Cultures ◽

Skin Physiology

In vitro culture of primary or established cell lines is one of the leading techniques in many areas of basic biological research. The use of pure or highly enriched cultures of specific cell types obtained from different tissues and genetics backgrounds has greatly contributed to our current understanding of normal and pathological cellular processes. Cells in culture are easily propagated generating an almost endless source of material for experimentation. Besides, they can be manipulated to achieve gene silencing, gene overexpression and genome editing turning possible the dissection of specific gene functions and signaling pathways. However, monolayer and suspension cultures of cells do not reproduce the cell type diversity, cell-cell contacts, cell-matrix interactions and differentiation pathways typical of the three-dimensional environment of tissues and organs from where they were originated. Therefore, different experimental animal models have been developed and applied to address these and other complex issues in vivo. However, these systems are costly and time consuming. Most importantly the use of animals in scientific research poses moral and ethical concerns facing a steadily increasing opposition from different sectors of the society. Therefore, there is an urgent need for the development of alternative in vitro experimental models that accurately reproduce the events observed in vivo to reduce the use of animals. Organotypic cultures combine the flexibility of traditional culture systems with the possibility of culturing different cell types in a 3D environment that reproduces both the structure and the physiology of the parental organ. Here we present a summarized description of the use of epithelial organotypic for the study of skin physiology, human papillomavirus biology and associated tumorigenesis.

Download Full-text

Hybrid Design of Isonicotinic Acid Hydrazide Derivatives: Machine Learning Studies, Synthesis and Biological Evaluation of their Antituberculosis Activity

Current Drug Discovery Technologies ◽

10.2174/1570163816666190411110331 ◽

2020 ◽

Vol 17 (3) ◽

pp. 365-375

Author(s):

Vasyl Kovalishyn ◽

Diana Hodyna ◽

Vitaliy O. Sinenko ◽

Volodymyr Blagodatny ◽

Ivan Semenyuta ◽

...

Keyword(s):

Machine Learning ◽

Isonicotinic Acid ◽

Acid Hydrazide ◽

Predictive Ability ◽

Antitubercular Activity ◽

Biological Evaluation ◽

Starting Point ◽

Binary Classifiers ◽

Synthesis And Biological Evaluation

Background: Tuberculosis (TB) is an infection disease caused by Mycobacterium tuberculosis (Mtb) bacteria. One of the main causes of mortality from TB is the problem of Mtb resistance to known drugs. Objective: The goal of this work is to identify potent small molecule anti-TB agents by machine learning, synthesis and biological evaluation. Methods: The On-line Chemical Database and Modeling Environment (OCHEM) was used to build predictive machine learning models. Seven compounds were synthesized and tested in vitro for their antitubercular activity against H37Rv and resistant Mtb strains. Results: A set of predictive models was built with OCHEM based on a set of previously synthesized isoniazid (INH) derivatives containing a thiazole core and tested against Mtb. The predictive ability of the models was tested by a 5-fold cross-validation, and resulted in balanced accuracies (BA) of 61–78% for the binary classifiers. Test set validation showed that the models could be instrumental in predicting anti- TB activity with a reasonable accuracy (with BA = 67–79 %) within the applicability domain. Seven designed compounds were synthesized and demonstrated activity against both the H37Rv and multidrugresistant (MDR) Mtb strains resistant to rifampicin and isoniazid. According to the acute toxicity evaluation in Daphnia magna neonates, six compounds were classified as moderately toxic (LD50 in the range of 10−100 mg/L) and one as practically harmless (LD50 in the range of 100−1000 mg/L). Conclusion: The newly identified compounds may represent a starting point for further development of therapies against Mtb. The developed models are available online at OCHEM http://ochem.eu/article/11 1066 and can be used to virtually screen for potential compounds with anti-TB activity.

Download Full-text

NKL Homeobox Gene VENTX Is Part of a Regulatory Network in Human Conventional Dendritic Cells

International Journal of Molecular Sciences ◽

10.3390/ijms22115902 ◽

2021 ◽

Vol 22 (11) ◽

pp. 5902

Author(s):

Stefan Nagel ◽

Claudia Pommerenke ◽

Corinna Meyer ◽

Hans G. Drexler

Keyword(s):

Dendritic Cells ◽

Transcription Factors ◽

Cell Lines ◽

Expression Profiling ◽

Regulatory Network ◽

Homeobox Gene ◽

Leukemia Cell Line ◽

Specific Gene ◽

Rna Seq ◽

And Function

Recently, we documented a hematopoietic NKL-code mapping physiological expression patterns of NKL homeobox genes in human myelopoiesis including monocytes and their derived dendritic cells (DCs). Here, we enlarge this map to include normal NKL homeobox gene expressions in progenitor-derived DCs. Analysis of public gene expression profiling and RNA-seq datasets containing plasmacytoid and conventional dendritic cells (pDC and cDC) demonstrated HHEX activity in both entities while cDCs additionally expressed VENTX. The consequent aim of our study was to examine regulation and function of VENTX in DCs. We compared profiling data of VENTX-positive cDC and monocytes with VENTX-negative pDC and common myeloid progenitor entities and revealed several differentially expressed genes encoding transcription factors and pathway components, representing potential VENTX regulators. Screening of RNA-seq data for 100 leukemia/lymphoma cell lines identified prominent VENTX expression in an acute myelomonocytic leukemia cell line, MUTZ-3 containing inv(3)(q21q26) and t(12;22)(p13;q11) and representing a model for DC differentiation studies. Furthermore, extended gene analyses indicated that MUTZ-3 is associated with the subtype cDC2. In addition to analysis of public chromatin immune-precipitation data, subsequent knockdown experiments and modulations of signaling pathways in MUTZ-3 and control cell lines confirmed identified candidate transcription factors CEBPB, ETV6, EVI1, GATA2, IRF2, MN1, SPIB, and SPI1 and the CSF-, NOTCH-, and TNFa-pathways as VENTX regulators. Live-cell imaging analyses of MUTZ-3 cells treated for VENTX knockdown excluded impacts on apoptosis or induced alteration of differentiation-associated cell morphology. In contrast, target gene analysis performed by expression profiling of knockdown-treated MUTZ-3 cells revealed VENTX-mediated activation of several cDC-specific genes including CSFR1, EGR2, and MIR10A and inhibition of pDC-specific genes like RUNX2. Taken together, we added NKL homeobox gene activities for progenitor-derived DCs to the NKL-code, showing that VENTX is expressed in cDCs but not in pDCs and forms part of a cDC-specific gene regulatory network operating in DC differentiation and function.

Download Full-text