scholarly journals Identifying transcriptomic correlates of histology using deep learning

PLoS ONE ◽  
2020 ◽  
Vol 15 (11) ◽  
pp. e0242858
Author(s):  
Liviu Badea ◽  
Emil Stănescu
Keyword(s):  
Gene Expression ◽  
Deep Learning ◽  
Expression Profiles ◽  
Expression Patterns ◽  
Gene Expression Profiles ◽  
Specific Gene ◽  
Genome Wide ◽  
Gene Perturbations ◽  
Whole Slide Images

Linking phenotypes to specific gene expression profiles is an extremely important problem in biology, which has been approached mainly by correlation methods or, more fundamentally, by studying the effects of gene perturbations. However, genome-wide perturbations involve extensive experimental efforts, which may be prohibitive for certain organisms. On the other hand, the characterization of the various phenotypes frequently requires an expert’s subjective interpretation, such as a histopathologist’s description of tissue slide images in terms of complex visual features (e.g. ‘acinar structures’). In this paper, we use Deep Learning to eliminate the inherent subjective nature of these visual histological features and link them to genomic data, thus establishing a more precisely quantifiable correlation between transcriptomes and phenotypes. Using a dataset of whole slide images with matching gene expression data from 39 normal tissue types, we first developed a Deep Learning tissue classifier with an accuracy of 94%. Then we searched for genes whose expression correlates with features inferred by the classifier and demonstrate that Deep Learning can automatically derive visual (phenotypical) features that are well correlated with the transcriptome and therefore biologically interpretable. As we are particularly concerned with interpretability and explainability of the inferred histological models, we also develop visualizations of the inferred features and compare them with gene expression patterns determined by immunohistochemistry. This can be viewed as a first step toward bridging the gap between the level of genes and the cellular organization of tissues.

scholarly journals Identifying Transcriptomic Correlates of Histology using Deep Learning

2020 ◽  
Author(s):  
Liviu Badea ◽  
Emil Stănescu
Keyword(s):  
Gene Expression ◽  
Deep Learning ◽  
Expression Profiles ◽  
Expression Patterns ◽  
Gene Expression Profiles ◽  
Specific Gene ◽  
Genome Wide ◽  
Gene Perturbations ◽  
Whole Slide Images

AbstractLinking phenotypes to specific gene expression profiles is an extremely important problem in biology, which has been approached mainly by correlation methods or, more fundamentally, by studying the effects of gene perturbations. However, genome-wide perturbations involve extensive experimental efforts, which may be prohibitive for certain organisms. On the other hand, the characterization of the various phenotypes frequently requires an expert’s subjective interpretation, such as a histopathologist’s description of tissue slide images in terms of complex visual features (e.g. ‘acinar structures’). In this paper, we use Deep Learning to eliminate the inherent subjective nature of these visual histological features and link them to genomic data, thus establishing a more precisely quantifiable correlation between transcriptomes and phenotypes. Using a dataset of whole slide images with matching gene expression data from 39 normal tissue types, we first developed a Deep Learning tissue classifier with an accuracy of 94%. Then we searched for genes whose expression correlates with features inferred by the classifier and demonstrate that Deep Learning can automatically derive visual (phenotypical) features that are well correlated with the transcriptome and therefore biologically interpretable. As we are particularly concerned with interpretability and explainability of the inferred histological models, we also develop visualizations of the inferred features and compare them with gene expression patterns determined by immunohistochemistry. This can be viewed as a first step toward bridging the gap between the level of genes and the cellular organization of tissues.

scholarly journals Interferon-regulated genetic programs and JAK/STAT pathway activate the intronic promoter of the short ACE2 isoform in renal proximal tubules

2021 ◽  
Author(s):  
Jakub Jankowski ◽  
Hye Kyung Lee ◽  
Julia Wilflingseder ◽  
Lothar Hennighausen
Keyword(s):  
Gene Expression ◽  
Proximal Tubule ◽  
Expression Profiles ◽  
Expression Patterns ◽  
Gene Expression Profiles ◽  
Regulatory Elements ◽  
Rna Seq ◽  
Angiotensin Converting Enzyme 2 ◽  
Genome Wide ◽  
Cytokine Stimulation

SummaryRecently, a short, interferon-inducible isoform of Angiotensin-Converting Enzyme 2 (ACE2), dACE2 was identified. ACE2 is a SARS-Cov-2 receptor and changes in its renal expression have been linked to several human nephropathies. These changes were never analyzed in context of dACE2, as its expression was not investigated in the kidney. We used Human Primary Proximal Tubule (HPPT) cells to show genome-wide gene expression patterns after cytokine stimulation, with emphasis on the ACE2/dACE2 locus. Putative regulatory elements controlling dACE2 expression were identified using ChIP-seq and RNA-seq. qRT-PCR differentiating between ACE2 and dACE2 revealed 300- and 600-fold upregulation of dACE2 by IFNα and IFNβ, respectively, while full length ACE2 expression was almost unchanged. JAK inhibitor ruxolitinib ablated STAT1 and dACE2 expression after interferon treatment. Finally, with RNA-seq, we identified a set of genes, largely immune-related, induced by cytokine treatment. These gene expression profiles provide new insights into cytokine response of proximal tubule cells.

scholarly journals Identification of Molecular Subgroups in Liver Cirrhosis by Gene Expression Profiles

2021 ◽  
Author(s):  
Ying-xue Zhang ◽  
Feng-xia Sun ◽  
Xiao-ling Li ◽  
Qing-hua Liu ◽  
Zi-meng Chen ◽  
...  
Keyword(s):  
Gene Expression ◽  
Liver Cirrhosis ◽  
Expression Analysis ◽  
Expression Profiles ◽  
Expression Patterns ◽  
Gene Expression Profiles ◽  
Receptor Interaction ◽  
Molecular Characteristics ◽  
Specific Gene ◽  
Subgroup Ii

Abstract Background: Cirrhosis is a common clinical chronic progressive liver disease and has become one of the main causes of death worldwide. The condition of liver cirrhosis is complex and there is also clinical heterogeneity. Identifying liver cirrhosis based on molecular characteristics has become a challenge.Methods: To reveal the potential molecular characteristics of different types of cirrhosis, we divided 79 patients with cirrhosis into 4 subgroups based on gene expression profiles. These gene expression profiles were retrieved from the mprehensive gene expression database. In addition, these subgroups showed different expression patterns. To reveal the differences between subgroups, we used weighted gene co-expression analysis and identified six subgroup-specific gene co-expression analysis modules.Results: The characteristics ofWCGNAmodules indicate that TGF - β signaling pathway,viral protein interaction with cytokines and cytokine receptors, including a variety of chemokines and inflammatory factors, are upregulated in subgroup I, indicating that subjects in subgroup I may show inflammatory characteristics; fatty acid metabolism, biosynthesis of cofactors, carbon metabolism and protein processing pathway in endoplasmic reticulum were significantly enriched in subgroup II, which indicated that the subjects in subgroup II might have the characteristics of active metabolism; arrhythmogenic right ventricular cardiomyopathy and Neuroactive ligand−receptor interaction are significantly enriched in subgroup IV; we did not find a significant upregulation pathway in the third subgroup.Conclusion: The subgroups classification of liver cirrhosis cases shows that patients from different subgroups may have unique gene expression patterns, which indicates that patients in each subgroup should receive more personalized treatment.

scholarly journals Meta-Analysis of RNA-Seq Studies Reveals Genes With Dominant Functions During Flower Bud Endo- To Eco-Dormancy Transition in Prunus Species

2021 ◽  
Author(s):  
Monica Canton ◽  
Cristian Forestan ◽  
Claudio Bonghi ◽  
Serena Varotto
Keyword(s):  
Gene Expression ◽  
Expression Profiles ◽  
Meta Analysis ◽  
Expression Patterns ◽  
Gene Expression Profiles ◽  
Fruit Trees ◽  
Day Length ◽  
Specific Gene ◽  
Flower Buds ◽  
Flower Bud

Abstract In deciduous fruit trees, entrance into dormancy occurs in later summer/fall, concomitantly with the shortening of day length and decrease in temperature. Dormancy can be divided into endodormancy, ecodormancy and paradormancy. In Prunus species flower buds, entrance into the dormant stage occurs when the apical meristem is partially differentiated; during dormancy, flower verticils continue their growth and differentiation. Each species and/or cultivar requires exposure to low winter temperature followed by warm temperatures, quantified as chilling and heat requirements, to remove the physiological blocks that inhibit budburst. A comprehensive meta-analysis of transcriptomic studies on flower buds of sweet cherry, apricot and peach was conducted, by investigating the gene expression profiles during bud endo- to ecodormancy transition in genotypes differing in chilling requirements. Conserved and distinctive expression patterns were observed, allowing the identification of gene specifically associated with endodormancy or ecodormancy. In addition to the MADS-box transcription factor family, hormone-related genes, chromatin modifiers, macro- and micro-gametogenesis related genes and environmental integrators, were identified as novel biomarker candidates for flower bud development during winter in stone fruits. In parallel, flower bud differentiation processes were associated to dormancy progression and termination and to environmental factors triggering dormancy phase-specific gene expression.

scholarly journals RNAlater and flash freezing storage methods nonrandomly influence observed gene expression in RNAseq experiments

2018 ◽  
Author(s):  
Courtney N. Passow ◽  
Thomas J. Y. Kono ◽  
Bethany A. Stahl ◽  
James B. Jaggard ◽  
Alex C. Keene ◽  
...  
Keyword(s):  
Gene Expression ◽  
Liquid Nitrogen ◽  
Expression Profiles ◽  
Expression Patterns ◽  
Gc Content ◽  
Gene Expression Profiles ◽  
Expression Studies ◽  
Genome Wide ◽  
Differential Gene ◽  
Genome Wide Gene Expression

AbstractRNA-sequencing is a popular next-generation sequencing technique for assaying genome-wide gene expression profiles. Nonetheless, it is susceptible to biases that are introduced by sample handling prior gene expression measurements. Two of the most common methods for preserving samples in both field-based and laboratory conditions are submersion in RNAlater and flash freezing in liquid nitrogen. Flash freezing in liquid nitrogen can be impractical, particularly for field collections. RNAlater is a solution for stabilizing tissue for longer-term storage as it rapidly permeates tissue to protect cellular RNA. In this study, we assessed genome-wide expression patterns in 30 day old fry collected from the same brood at the same time point that were flash-frozen in liquid nitrogen and stored at −80°C or submerged and stored in RNAlater at room temperature, simulating conditions of fieldwork. We show that sample storage is a significant factor influencing observed differential gene expression. In particular, genes with elevated GC content exhibit higher observed expression levels in liquid nitrogen flash-freezing relative to RNAlater-storage. Further, genes with higher expression in RNAlater relative to liquid nitrogen experience disproportionate enrichment for functional categories, many of which are involved in RNA processing. This suggests that RNAlater may elicit a physiological response that has the potential to bias biological interpretations of expression studies. The biases introduced to observed gene expression arising from mimicking many field-based studies are substantial and should not be ignored.

scholarly journals Boosting Gene Expression Clustering with System-Wide Biological Information: A Robust Autoencoder Approach

2017 ◽  
Author(s):  
Hongzhu Cui ◽  
Chong Zhou ◽  
Xinyu Dai ◽  
Yuting Liang ◽  
Randy Paffenroth ◽  
...  
Keyword(s):  
Gene Expression ◽  
Deep Learning ◽  
Expression Profiles ◽  
Gene Expression Profiles ◽  
Gene Expression Dataset ◽  
Biological Information ◽  
Clustering Methods ◽  
Genome Wide ◽  
A Cell ◽  
High Level

AbstractGene expression analysis provides genome-wide insights into the transcriptional activity of a cell. One of the first computational steps in exploration and analysis of the gene expression data is clustering. With a number of standard clustering methods routinely used, most of the methods do not take prior biological information into account. In this paper, we propose a new approach for gene expression clustering analysis. The approach benefits from a new deep learning architecture, Robust Autoencoder, which provides a more accurate high-level representation of the feature sets, and from incorporating prior biological information into the clustering process. We tested our approach on two distinct gene expression datasets and compared the performance with two widely used clustering methods, hierarchical clustering and k-means, as well as with a recent deep learning clustering approach. As a result, our approach outperformed all other clustering methods on the labeled yeast gene expression dataset. Furthermore we showed that it is better in identifying the functionally common clusters than k-means on the unlabeled human gene expression dataset. The results demonstrate that our new deep learning architecture could generalize well the specific properties of gene expression profiles. Furthermore, the results confirm our hypothesis that the prior biological network knowledge could be helpful in the gene expression clustering task.

Automated Sorting of Monocytes from Whole Blood for Reproducible Gene Expression Profiling.

Blood ◽  
2007 ◽  
Vol 110 (11) ◽  
pp. 3840-3840
Author(s):  
Carsten Poggel ◽  
Timo Adams ◽  
Sabine Martin ◽  
Carola Pickel ◽  
Nicole Prahl ◽  
...  
Keyword(s):  
Gene Expression ◽  
Blood Cell ◽  
Gene Expression Profiling ◽  
Whole Blood ◽  
Expression Profiling ◽  
Expression Profiles ◽  
Expression Patterns ◽  
Gene Expression Profiles ◽  
Cell Types ◽  
Specific Gene

Abstract Microarray-based gene expression profiling has been used to develop clinically relevant molecular classifiers for many different diseases. Furthermore, it has been shown for various chronic diseases that specific gene expression patterns are reflected at the level of blood cells. However, blood is a complex tissue comprising numerous cell types. Therefore, the contribution of rare cell types to a whole blood expression profile might not be detected and a substantial proportion of what is usually reported as “up-regulation” or “down-regulation” might actually be the result of a shift in cell populations and not of a true regulatory process. In order to circumvent these problems, several techniques have been established to analyze purified subpopulations rather than whole blood samples. Previously, it has been shown, for example, that reproducible gene expression profiles can be generated by positive selection of blood cell subsets from PBMCs1. As the preparation of PBMCs by, for example, Ficoll is time-consuming, inconvenient, and not amenable to automation, we have set up a combined direct whole blood cell separation and gene expression profiling protocol. By using Whole Blood CD14 MicroBeads in combination with the autoMACS Pro™ Separator, the separation protocol generally allowed enrichment of monocytes from whole blood within 30 min with purities higher than 90%. In combination with the depletion of neutrophils, the major source of contaminating RNA, purities increased to over 95% for all tested blood donors. Monocytes included the CD14bright/CD16− as well as the CD14dim/CD16+ populations. To assess the reproducibility of gene expression profiles and the influence of several experimental parameters, monocytes were sorted from 5 ml whole blood. RNA was extracted and hybridized to microarrays and the Pearson correlation coefficients of pairwise comparisons were calculated. Technical repeats of monocyte analysis from blood donated at different days showed a higher correlation coefficient than whole blood RNA. Blood storage at room temperature resulted in a strong deregulation of many genes, whereas blood stored at 4°C showed minimal changes, which is in agreement with previous studies. Skipping the centrifugation step, which is used to remove unbound MicroBeads did not alter the gene expression profiles. Incubation of sorted cells in PrepProtect™ Stabilization Buffer showed no alteration of gene expression thus enabling the shipping of cells without liquid nitrogen. Monocytes play a crucial role in diseases like atherosclerosis. Our rapid and simple protocol for combined direct cell sorting from whole blood and gene expression profiling of monocytes might help to ease the discovery of new biomarkers and to screen and monitor patients. 1 Lyons et al., BMC Genomics (2007), 8:64.

Characterization of intestinal gene expression profiles in Crohnʼs disease by genome-wide microarray analysis

2010 ◽  
Vol 16 (10) ◽  
pp. 1717-1728 ◽  
Author(s):  
Colin L. Noble ◽  
Alexander R. Abbas ◽  
Charles W. Lees ◽  
Jennine Cornelius ◽  
Karen Toy ◽  
...  
Keyword(s):  
Gene Expression ◽  
Microarray Analysis ◽  
Expression Profiles ◽  
Gene Expression Profiles ◽  
Genome Wide

Tumor-specific gene expression patterns with gene expression profiles

2006 ◽  
Vol 49 (3) ◽  
pp. 293-304 ◽  
Author(s):  
Xiaogang Ruan ◽  
Yingxin Li ◽  
Jiangeng Li ◽  
Daoxiong Gong ◽  
Jinlian Wang
Keyword(s):  
Gene Expression ◽  
Expression Profiles ◽  
Expression Patterns ◽  
Gene Expression Profiles ◽  
Gene Expression Patterns ◽  
Specific Gene ◽  
Specific Gene Expression

scholarly journals Meta-analysis of RNA-Seq studies reveals genes with dominant functions during flower bud endo- to eco-dormancy transition in Prunus species

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Monica Canton ◽  
Cristian Forestan ◽  
Claudio Bonghi ◽  
Serena Varotto
Keyword(s):  
Gene Expression ◽  
Expression Profiles ◽  
Meta Analysis ◽  
Expression Patterns ◽  
Gene Expression Profiles ◽  
Fruit Trees ◽  
Day Length ◽  
Specific Gene ◽  
Flower Buds ◽  
Flower Bud

AbstractIn deciduous fruit trees, entrance into dormancy occurs in later summer/fall, concomitantly with the shortening of day length and decrease in temperature. Dormancy can be divided into endodormancy, ecodormancy and paradormancy. In Prunus species flower buds, entrance into the dormant stage occurs when the apical meristem is partially differentiated; during dormancy, flower verticils continue their growth and differentiation. Each species and/or cultivar requires exposure to low winter temperature followed by warm temperatures, quantified as chilling and heat requirements, to remove the physiological blocks that inhibit budburst. A comprehensive meta-analysis of transcriptomic studies on flower buds of sweet cherry, apricot and peach was conducted, by investigating the gene expression profiles during bud endo- to ecodormancy transition in genotypes differing in chilling requirements. Conserved and distinctive expression patterns were observed, allowing the identification of gene specifically associated with endodormancy or ecodormancy. In addition to the MADS-box transcription factor family, hormone-related genes, chromatin modifiers, macro- and micro-gametogenesis related genes and environmental integrators, were identified as novel biomarker candidates for flower bud development during winter in stone fruits. In parallel, flower bud differentiation processes were associated to dormancy progression and termination and to environmental factors triggering dormancy phase-specific gene expression.

Sign in / Sign up

Export Citation Format

Share Document