Computational deconvolution of transcriptomics data from mixed cell populations

Niche deconvolution of the glioblastoma proteome reveals a distinct infiltrative phenotype within the proneural transcriptomic subgroup

Glioblastoma is often subdivided into three transcriptional subtypes (classical, proneural, mesenchymal) based on bulk RNA signatures that correlate with distinct genetic and clinical features. Potential cellular-level differences of these subgroups, such as the relative proportions of glioblastoma’s hallmark histopathologic features (e.g. brain infiltration, microvascular proliferation), may provide insight into their distinct phenotypes but are, however, not well understood. Here we leverage machine learning and reference proteomic profiles derived from micro-dissected samples of these major histomorphologic glioblastoma features to deconvolute and estimate niche proportions in an independent proteogenomically-characterized cohort. This approach revealed a strong association of the proneural transcriptional subtype with a diffusely infiltrating phenotype. Similarly, enrichment of a microvascular proliferation proteomic signature was seen within the mesenchymal subtype. This study is the first to link differences in the cellular pathology signatures and transcriptional profiles of glioblastoma, providing potential new insights into the genetic drivers and poor treatment response of specific subsets of glioblastomas.

Early transcriptional responses of bronchial epithelial cells to whole cigarette smoke mirror those of in-vivo exposed human bronchial mucosa

BACKGROUND

Despite the well-known detrimental effects of cigarette smoke (CS), little is known about the complex gene expression dynamics in the early stages after exposure. This study aims to investigate early transcriptomic responses following CS exposure of airway epithelial cells in culture and compare these to those found in human CS exposure studies.

METHODS

Primary bronchial epithelial cells (PBEC) were differentiated at the air-liquid interface (ALI) and exposed to whole CS. Bulk RNA-sequencing was performed at 1 h, 4 h, and 24 h hereafter, followed by differential gene expression analysis. Results were additionally compared to data retrieved from human CS studies.

RESULTS

ALI-PBEC gene expression in response to CS was most significantly changed at 4 h after exposure. Early transcriptomic changes (1 h, 4 h post CS exposure) were related to oxidative stress, xenobiotic metabolism, higher expression of immediate early genes and pro-inflammatory pathways (i.e., Nrf2, AP-1, AhR). At 24 h, ferroptosis-associated genes were significantly increased, whereas PRKN, involved in removing dysfunctional mitochondria, was downregulated. Importantly, the transcriptome dynamics of the current study mirrored in-vivo human studies of acute CS exposure, chronic smokers, and inversely mirrored smoking cessation.

CONCLUSION

These findings show that early after CS exposure xenobiotic metabolism and pro-inflammatory pathways were activated, followed by activation of the ferroptosis-related cell death pathway. Moreover, significant overlap between these transcriptomic responses in the in-vitro model and human in-vivo studies was found, with an early response of ciliated cells. These results provide validation for the use of ALI-PBEC cultures to study the human lung epithelial response to inhaled toxicants.

GEOMETRIC STRUCTURE GUIDED MODEL AND ALGORITHMS FOR COMPLETE DECONVOLUTION OF GENE EXPRESSION DATA

Complete deconvolution analysis for bulk RNA-seq data is important and helpful to distinguish whether the differences of disease-associated GEPs (gene expression profiles) in tissues of patients and normal controls are due to changes in cellular composition of tissue samples, or due to GEPs changes in specific cells. One of the major techniques to perform complete deconvolution is nonnegative matrix factorization (NMF), which also has a wide-range of applications in the machine learning community. However, the NMF is a well-known strongly ill-posed problem, so a direct application of NMF to RNA-seq data will suffer severe difficulties in the interpretability of solutions. In this paper, we develop an NMF-based mathematical model and corresponding computational algorithms to improve the solution identifiability of deconvoluting bulk RNA-seq data. In our approach, we combine the biological concept of marker genes with the solvability conditions of the NMF theories, and develop a geometric structures guided optimization model. In this strategy, the geometric structure of bulk tissue data is first explored by the spectral clustering technique. Then, the identified information of marker genes is integrated as solvability constraints, while the overall correlation graph is used as manifold regularization. Both synthetic and biological data are used to validate the proposed model and algorithms, from which solution interpretability and accuracy are significantly improved.

Integrative analysis of drug response and clinical outcome in acute myeloid leukemia

Acute myeloid leukemia (AML) is a cancer of myeloid-lineage cells with limited therapeutic options. We previously combined ex vivo drug sensitivity with genomic, transcriptomic, and clinical annotations for a large cohort of AML patients, which facilitated discovery of functional genomic correlates. Here, we present a dataset that has been harmonized with our initial report to yield a cumulative cohort of 805 patients (942 specimens). We show strong cross-cohort concordance and identify features of drug response. Further, deconvoluting transcriptomic data shows that drug sensitivity is governed broadly by AML cell differentiation state, sometimes conditionally affecting other correlates of response. Finally, modeling of clinical outcome reveals a single gene, PEAR1, to be among the strongest predictors of patient survival, especially for young patients. Collectively, this report expands a large functional genomic resource, offers avenues for mechanistic exploration and drug development, and reveals tools for predicting outcome in AML.

Future directions for precision oncology in prostate cancer

Clinical genomic testing is becoming routine in prostate cancer, as biomarker-driven therapies such as poly-ADP ribose polymerase (PARP) inhibitors and anti-PD1 immunotherapy are now approved for select men with castration-resistant prostate cancer harboring alterations in DNA repair genes. Challenges for precision medicine in prostate cancer include an overall low prevalence of actionable genomic alterations and a still limited understanding of the impact of tumor heterogeneity and co-occurring alterations on treatment response and outcomes across diverse patient populations. Expanded tissue-based technologies such as whole-genome sequencing, transcriptome analysis, epigenetic analysis, and single-cell RNA sequencing have not yet entered the clinical realm and could potentially improve upon our understanding of how molecular features of tumors, intratumoral heterogeneity, and the tumor microenvironment impact therapy response and resistance. Blood-based technologies including cell-free DNA, circulating tumor cells (CTCs), and extracellular vesicles (EVs) are less invasive molecular profiling resources that could also help capture intraindividual tumor heterogeneity and track dynamic changes that occur in the context of specific therapies. Furthermore, molecular imaging is an important biomarker tool within the framework of prostate cancer precision medicine with a capability to detect heterogeneity across metastases and potential therapeutic targets less invasively. Here, we review recent technological advances that may help promote the future implementation and value of precision oncology testing for patients with advanced prostate cancer.

Spatial Transcriptomics as a Novel Approach to Redefine Electrical Stimulation Safety

Current standards for safe delivery of electrical stimulation to the central nervous system are based on foundational studies which examined post-mortem tissue for histological signs of damage. This set of observations and the subsequently proposed limits to safe stimulation, termed the "Shannon limits," allow for a simple calculation (using charge per phase and charge density) to determine the intensity of electrical stimulation that can be delivered safely to brain tissue. In the three decades since the Shannon limits were reported, advances in molecular biology have allowed for more nuanced and detailed approaches to be used to expand current understanding of the physiological effects of stimulation. Here, we demonstrate the use of spatial transcriptomics (ST) in an exploratory investigation to assess the biological response to electrical stimulation in the brain. Electrical stimulation was delivered to the rat visual cortex with either acute or chronic electrode implantation procedures. To explore the influence of device type and stimulation parameters, we used carbon fiber ultramicroelectrode arrays (7 μm diameter) and microwire electrode arrays (50 μm diameter) delivering charge and charge density levels selected above and below reported tissue damage thresholds (range: 2-20 nC, 0.1-1 mC/cm2). Spatial transcriptomics was performed using Visium Spatial Gene Expression Slides (10x Genomics, Pleasanton, CA, United States), which enabled simultaneous immunohistochemistry and ST to directly compare traditional histological metrics to transcriptional profiles within each tissue sample. Our data give a first look at unique spatial patterns of gene expression that are related to cellular processes including inflammation, cell cycle progression, and neuronal plasticity. At the acute timepoint, an increase in inflammatory and plasticity related genes was observed surrounding a stimulating electrode compared to a craniotomy control. At the chronic timepoint, an increase in inflammatory and cell cycle progression related genes was observed both in the stimulating vs. non-stimulating microwire electrode comparison and in the stimulating microwire vs. carbon fiber comparison. Using the spatial aspect of this method as well as the within-sample link to traditional metrics of tissue damage, we demonstrate how these data may be analyzed and used to generate new hypotheses and inform safety standards for stimulation in cortex.

Deconvolution of a Large Cohort of Placental Microarray Data Reveals Clinically Distinct Subtypes of Preeclampsia

It has been well established that the dysfunctional placenta plays an important role in the pathogenesis of preeclampsia (PE), a hypertensive disorder in pregnancy. However, it is not well understood how individual cell types in the placenta are involved in placenta dysfunction because of limited single-cell studies of placenta with PE. Given that a high-resolution single-cell atlas in the placenta is now available, deconvolution of publicly available bulk PE transcriptome data may provide us with the opportunity to investigate the contribution of individual placental cell types to PE. Recent benchmark studies on deconvolution have provided suggestions on the strategy of marker gene selection and the choice of methodologies. In this study, we experimented with these suggestions by using real bulk data with known cell-type proportions and established a deconvolution pipeline using CIBERSORT. Applying the deconvolution pipeline to a large cohort of PE placental microarray data, we found that the proportions of trophoblast cells in the placenta were significantly different between PE and normal controls. We then predicted cell-type-level expression profiles for each sample using CIBERSORTx and found that the activities of several canonical PE-related pathways were significantly altered in specific subtypes of trophoblasts in PE. Finally, we constructed an integrated expression profile for each PE sample by combining the predicted cell-type-level expression profiles of several clinically relevant placental cell types and identified four clusters likely representing four PE subtypes with clinically distinct features. As such, our study showed that deconvolution of a large cohort of placental microarray provided new insights about the molecular mechanism of PE that would not be obtained by analyzing bulk expression profiles.

Distinct cell type-specific protein signatures in GRN and MAPT genetic subtypes of frontotemporal dementia

Frontotemporal dementia is characterized by progressive atrophy of frontal and/or temporal cortices at an early age of onset. The disorder shows considerable clinical, pathological, and genetic heterogeneity. Here we investigated the proteomic signatures of frontal and temporal cortex from brains with frontotemporal dementia due to GRN and MAPT mutations to identify the key cell types and molecular pathways in their pathophysiology. We compared patients with mutations in the GRN gene (n = 9) or with mutations in the MAPT gene (n = 13) with non-demented controls (n = 11). Using quantitative proteomic analysis on laser-dissected tissues we identified brain region-specific protein signatures for both genetic subtypes. Using published single cell RNA expression data resources we deduced the involvement of major brain cell types in driving these different protein signatures. Subsequent gene ontology analysis identified distinct genetic subtype- and cell type-specific biological processes. For the GRN subtype, we observed a distinct role for immune processes related to endothelial cells and for mitochondrial dysregulation in neurons. For the MAPT subtype, we observed distinct involvement of dysregulated RNA processing, oligodendrocyte dysfunction, and axonal impairments. Comparison with an in-house protein signature of Alzheimer’s disease brains indicated that the observed alterations in RNA processing and oligodendrocyte function are distinct for the frontotemporal dementia MAPT subtype. Taken together, our results indicate the involvement of different brain cell types and biological mechanisms in genetic subtypes of frontotemporal dementia. Furthermore, we demonstrate that comparison of proteomic profiles of different disease entities can separate general neurodegenerative processes from disease-specific pathways, which may aid the development of disease subtype-specific treatment strategies.

Predicting Algorithm of Tissue Cell Ratio Based on Deep Learning Using Single-Cell RNA Sequencing

Background: Understanding the proportion of cell types in heterogeneous tissue samples is important in bioinformatics. It is a challenge to infer the proportion of tissues using bulk RNA sequencing data in bioinformatics because most traditional algorithms for predicting tissue cell ratios heavily rely on standardized specific cell-type gene expression profiles, and do not consider tissue heterogeneity. The prediction accuracy of algorithms is limited, and robustness is lacking. This means that new approaches are needed urgently. Methods: In this study, we introduced an algorithm that automatically predicts tissue cell ratios named Autoptcr. The algorithm uses the data simulated by single-cell RNA sequencing (ScRNA-Seq) for model training, using convolutional neural networks (CNNs) to extract intrinsic relationships between genes and predict the cell proportions of tissues. Results: We trained the algorithm using simulated bulk samples and made predictions using real bulk PBMC data. Comparing Autoptcr with existing advanced algorithms, the Pearson correlation coefficient between the actual value of Autoptcr and the predicted value was the highest, reaching 0.903. Tested on a bulk sample, the correlation coefficient of Lin was 41% higher than that of CSx. The algorithm can infer tissue cell proportions directly from tissue gene expression data. Conclusions: The Autoptcr algorithm uses simulated ScRNA-Seq data for training to solve the problem of specific cell-type gene expression profiles. It also has high prediction accuracy and strong noise resistance for the tissue cell ratio. This work is expected to provide new research ideas for the prediction of tissue cell proportions.

A novel integrative computational framework for breast cancer radiogenomic biomarker discovery

•

Bayesian tensor factorization is used to integrate multiomics data for radiogenomics analysis.

•

A regression framework is proposed to handle the unmatched data issue in radiogenomics analysis.

•

Deep learning is used to identify prognostic meaningful radiogenomic biomarkers for cancer.

In precise medicine, it is with great value to develop computational frameworks for identifying prognostic biomarkers which can capture both multi-genomic and phenotypic heterogeneity of breast cancer (BC). Radiogenomics is a field where medical images and genomic measurements are integrated and mined to solve challenging clinical problems. Previous radiogenomic studies suffered from data incompleteness, feature subjectivity and low interpretability. For example, the majority of the radiogenomic studies miss one or two of medical imaging data, genomic data, and clinical outcome data, which results in the data incomplete issue. Feature subjectivity issue comes from the extraction of imaging features with significant human involvement. Thus, there is an urgent need to address above-mentioned limitations so that fully automatic and transparent radiogenomic prognostic biomarkers could be identified for BC.

We proposed a novel framework for BC prognostic radiogenomic biomarker identification. This framework involves an explainable DL model for image feature extraction, a Bayesian tensor factorization (BTF) processing for multi-genomic feature extraction, a leverage strategy to utilize unpaired imaging, genomic, and survival outcome data, and a mediation analysis to provide further interpretation for identified biomarkers. This work provided a new perspective for conducting a comprehensive radiogenomic study when only limited resources are given. Compared with baseline traditional radiogenomic biomarkers, the 23 biomarkers identified by the proposed framework performed better in indicating patients’ survival outcome. And their interpretability is guaranteed by different levels of build-in and follow-up analyses.

Coding and non-coding co-expression network analysis identifies key modules and driver genes associated with precursor lesions of gastric cancer

BACKGROUND

Helicobacter pylori infection is the most important risk factor for gastric cancer (GC). Human gastric adenocarcinoma develops after long-term H. pylori infection via the Correa cascade. This carcinogenic pathway describes the progression from gastritis to atrophy, intestinal metaplasia (IM), dysplasia and GC. Patients with atrophy and intestinal metaplasia are considered to have precancerous lesions of GC (PLGC). H. pylori eradication and endoscopy surveillance are currently the only interventions for preventing GC. Better knowledge of the biology of human PLGC may help find stratification markers and contribute to better understanding of biological mechanisms. One way to achieve this is by using co-expression network analysis. Weighted gene co-expression network analysis (WGCNA) is often used to identify modules from co-expression networks and relate them to clinical traits. It also allows identification of driver genes that may be critical for PLGC.

AIM

The purpose of this study was to identify co-expression modules and differential gene expression in dyspeptic patients at different stages of the Correa pathway.

METHODS

We studied 96 gastric biopsies from 78 patients that were clinically classified as: non-active (n = 10) and chronic-active gastritis (n = 20), atrophy (n = 12), and IM (n = 36). Gene expression of coding RNAs was determined by microarrays and non-coding RNAs by RNA-seq. The WGCNA package was used for network construction, module detection, module preservation and hub and driver gene selection.

RESULTS

WGCNA identified 20 modules for coding RNAs and 4 for each miRNA and small RNA class. Modules were associated with antrum and corpus gastric locations, chronic gastritis and IM. Notably, coding RNA modules correlated with the Correa cascade. One was associated with the presence of H. pylori. In three modules, the module eigengene (ME) gradually increased in the stages toward IM, while in three others the inverse relationship was found. One miRNA module was negatively correlated to IM and was used for a mRNA-miRNA integration analysis. WGCNA also uncovered driver genes. Driver genes show both high connectivity within a module and are significantly associated with clinical traits. Some of those genes have been previously involved in H. pylori carcinogenesis, but others are new. Lastly, using similar external transcriptomic data, we confirmed that the discovered mRNA modules were highly preserved.

CONCLUSION

Our analysis captured co-expression modules that provide valuable information to understand the pathogenesis of the progression of PLGC.

scDeconv: an R package to deconvolve bulk DNA methylation data with scRNA-seq data and paired bulk RNA-DNA methylation data

Many DNA methylation (DNAm) data are from tissues composed of various cell types, and hence cell deconvolution methods are needed to infer their cell compositions accurately. However, a bottleneck for DNAm data is the lack of cell-type-specific DNAm references. On the other hand, scRNA-seq data are being accumulated rapidly with various cell-type transcriptomic signatures characterized, and also, many paired bulk RNA-DNAm data are publicly available currently. Hence, we developed the R package scDeconv to use these resources to solve the reference deficiency problem of DNAm data and deconvolve them from scRNA-seq data in a trans-omics manner. It assumes that paired samples have similar cell compositions. So the cell content information deconvolved from the scRNA-seq and paired RNA data can be transferred to the paired DNAm samples. Then an ensemble model is trained to fit these cell contents with DNAm features and adjust the paired RNA deconvolution in a co-training manner. Finally, the model can be used on other bulk DNAm data to predict their relative cell-type abundances. The effectiveness of this method is proved by its accurate deconvolution on the three testing datasets here, and if given an appropriate paired dataset, scDeconv can also deconvolve other omics, such as ATAC-seq data. Furthermore, the package also contains other functions, such as identifying cell-type-specific inter-group differential features from bulk DNAm data. scDeconv is available at: https://github.com/yuabrahamliu/scDeconv.

Tumor Microenvironment Profiling Identifies Prognostic Signatures and Suggests Immunotherapeutic Benefits in Neuroblastoma

The tumor microenvironment (TME) influences disease initiation and progression. Cross-talks of cells within TME can affect the efficacy of immunotherapies. However, a precise, concise, and comprehensive TME landscape in neuroblastoma (NB) has not been established. Here, we profiled the TME landscape of 498 NB-related patients on a self-curated gene list and identified three prognostic TMEsubgroups. The differentially expressed genes in these three TMEsubgroups were used to construct a genetic signature of the TME landscape and characterize three GeneSubgroups. The subgroup with the worst overall survival prognosis, the TMEsubgroup/GeneSubgroup3, lacked immune cell infiltration and received the highest scores of MYCN- and ALK-related signatures and lowest scores of immune pathways. Additionally, we found that the GeneSubgroup3 might be benefited from anti-GD2 instead of anti-PD-1 therapy. We further created a 48-gene signature, the TMEscore, to infer prognosis and validated it in three independent NB cohorts and a pan-cancer cohort of 9,460 patients. We did RNA-seq on 16 samples and verified that TMEscore was higher in patients with stage 3/4 than stage 1/2 diseases. The TMEscore could also predict responses for several immunotherapies. After adding clinical features, we found that the nomogram-based score system, the TMEIndex, surpassed the current risk system at predicting survivals. Our analysis explained TME at the transcriptome level and paved the way for immunotherapies in NB.

The Effect of Hypoxia and Hypoxia-Associated Pathways in the Regulation of Antitumor Response: Friends or Foes?

Hypoxia is an environmental stressor that is instigated by low oxygen availability. It fuels the progression of solid tumors by driving tumor plasticity, heterogeneity, stemness and genomic instability. Hypoxia metabolically reprograms the tumor microenvironment (TME), adding insult to injury to the acidic, nutrient deprived and poorly vascularized conditions that act to dampen immune cell function. Through its impact on key cancer hallmarks and by creating a physical barrier conducive to tumor survival, hypoxia modulates tumor cell escape from the mounted immune response. The tumor cell-immune cell crosstalk in the context of a hypoxic TME tips the balance towards a cold and immunosuppressed microenvironment that is resistant to immune checkpoint inhibitors (ICI). Nonetheless, evidence is emerging that could make hypoxia an asset for improving response to ICI. Tackling the tumor immune contexture has taken on an in silico, digitalized approach with an increasing number of studies applying bioinformatics to deconvolute the cellular and non-cellular elements of the TME. Such approaches have additionally been combined with signature-based proxies of hypoxia to further dissect the turbulent hypoxia-immune relationship. In this review we will be highlighting the mechanisms by which hypoxia impacts immune cell functions and how that could translate to predicting response to immunotherapy in an era of machine learning and computational biology.

Comprehensive evaluation of deconvolution methods for human brain gene expression

Transcriptome deconvolution aims to estimate the cellular composition of an RNA sample from its gene expression data, which in turn can be used to correct for composition differences across samples. The human brain is unique in its transcriptomic diversity, and comprises a complex mixture of cell-types, including transcriptionally similar subtypes of neurons. Here, we carry out a comprehensive evaluation of deconvolution methods for human brain transcriptome data, and assess the tissue-specificity of our key observations by comparison with human pancreas and heart. We evaluate eight transcriptome deconvolution approaches and nine cell-type signatures, testing the accuracy of deconvolution using in silico mixtures of single-cell RNA-seq data, RNA mixtures, as well as nearly 2000 human brain samples. Our results identify the main factors that drive deconvolution accuracy for brain data, and highlight the importance of biological factors influencing cell-type signatures, such as brain region and in vitro cell culturing.

Transcriptome deconvolution aims to estimate cellular composition based on gene expression data. Here the authors evaluate deconvolution methods for human brain transcriptome and conclude that partial deconvolution algorithms work best, but that appropriate cell-type signatures are also important.

A protocol to extract cell-type-specific signatures from differentially expressed genes in bulk-tissue RNA-seq

Bulk-tissue RNA-seq is widely used to dissect variation in gene expression levels across tissues and under different experimental conditions. Here, we introduce a protocol that leverages existing single-cell expression data to deconvolve patterns of cell-type-specific gene expression in differentially expressed gene lists from highly heterogeneous tissue. We apply this protocol to interrogate cell-type-specific gene expression and variation in cell type composition between the distinct sublayers of the hippocampal CA1 region of the brain in a rodent model of epilepsy.

For complete details on the use and execution of this protocol, please refer to Cid et al. (2021).

•

A protocol to explore gene signatures from bulk RNA-seq at the cell-type-specific level

•

Deconvolution of complex gene signatures from highly heterogeneous tissues

•

Publicly available single-cell gene expression dataset is retrieved and curated

•

Gene signatures across brain regions and disease states are surveyed in scRNA-seq data

swCAM: estimation of subtype-specific expressions in individual samples with unsupervised sample-wise deconvolution

MOTIVATION

Complex biological tissues are often a heterogeneous mixture of several molecularly distinct cell subtypes. Both subtype compositions and subtype-specific (STS) expressions can vary across biological conditions. Computational deconvolution aims to dissect patterns of bulk tissue data into subtype compositions and STS expressions. Existing deconvolution methods can only estimate averaged STS expressions in a population, while many downstream analyses such as inferring co-expression networks in particular subtypes require subtype expression estimates in individual samples. However, individual-level deconvolution is a mathematically underdetermined problem because there are more variables than observations.

RESULTS

We report a sample-wise Convex Analysis of Mixtures (swCAM) method that can estimate subtype proportions and STS expressions in individual samples from bulk tissue transcriptomes. We extend our previous CAM framework to include a new term accounting for between-sample variations and formulate swCAM as a nuclear-norm and ℓ2,1-norm regularized matrix factorization problem. We determine hyperparameter values using cross-validation with random entry exclusion and obtain a swCAM solution using an efficient alternating direction method of multipliers. Experimental results on realistic simulation data show that swCAM can accurately estimate STS expressions in individual samples and successfully extract co-expression networks in particular subtypes that are otherwise unobtainable using bulk data. In two real-world applications, swCAM analysis of bulk RNASeq data from brain tissue of cases and controls with bipolar disorder or Alzheimer's disease identified significant changes in cell proportion, expression pattern and co-expression module in patient neurons. Comparative evaluation of swCAM versus peer methods is also provided.

AVAILABILITY AND IMPLEMENTATION

The R Scripts of swCAM are freely available at https://github.com/Lululuella/swCAM. A user's guide and a vignette are provided.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

DecOT: Bulk Deconvolution With Optimal Transport Loss Using a Single-Cell Reference

Tissues are constituted of heterogeneous cell types. Although single-cell RNA sequencing has paved the way to a deeper understanding of organismal cellular composition, the high cost and technical noise have prevented its wide application. As an alternative, computational deconvolution of bulk tissues can be a cost-effective solution. In this study, we propose DecOT, a deconvolution method that uses the Wasserstein distance as a loss and applies scRNA-seq data as references to characterize the cell type composition from bulk tissue RNA-seq data. The Wasserstein loss in DecOT is able to utilize additional information from gene space. DecOT also applies an ensemble framework to integrate deconvolution results from multiple individuals’ references to mitigate the individual/batch effect. By benchmarking DecOT with four recently proposed square loss-based methods on pseudo-bulk data from four different single-cell data sets and real pancreatic islet bulk samples, we show that DecOT outperforms other methods and the ensemble framework is robust to the choice of references.

Unraveling patterns of disrupted gene expression across a complex tissue

Whole tissue RNASeq is the standard approach for studying gene expression divergence in evolutionary biology and provides a snapshot of the comprehensive transcriptome for a given tissue. However, whole tissues consist of diverse cell types differing in expression profiles, and the cellular composition of these tissues can evolve across species. Here, we investigate the effects of different cellular composition on whole tissue expression profiles. We compared gene expression from whole testes and enriched spermatogenesis populations in two species of house mice, Mus musculus musculus and M. m. domesticus, and their sterile and fertile F1 hybrids, which differ in both cellular composition and regulatory dynamics. We found that cellular composition differences skewed expression profiles and differential gene expression in whole testes samples. Importantly, both approaches were able to detect large-scale patterns such as disrupted X chromosome expression, although whole testes sampling resulted in decreased power to detect differentially expressed genes. We encourage researchers to account for histology in RNASeq and consider methods that reduce sample complexity whenever feasible. Ultimately, we show that differences in cellular composition between tissues can modify expression profiles, potentially altering inferred gene ontological processes, insights into gene network evolution, and processes governing gene expression evolution.

A2Sign: Agnostic Algorithms for Signatures-a universal method for identifying molecular signatures from transcriptomic datasets prior to cell-type deconvolution

MOTIVATION

Molecular signatures are critical for inferring the proportions of cell types from bulk transcriptomics data. However, the identification of these signatures is based on a methodology that relies on prior biological knowledge of the cell types being studied. When working with less known biological material, a data-driven approach is required to uncover the underlying classes and generate ad hoc signatures from healthy or pathogenic tissue.

RESULTS

We present a new approach, A2Sign: Agnostic Algorithms for Signatures, based on a non-negative tensor factorization (NTF) strategy that allows us to identify cell-type-specific molecular signatures, greatly reduce collinearities and also account for inter-individual variability. We propose a global framework that can be applied to uncover molecular signatures for cell-type deconvolution in arbitrary tissues using bulk transcriptome data. We also present two new molecular signatures for deconvolution of up to 16 immune cell types using microarray or RNA-seq data.

AVAILABILITY AND IMPLEMENTATION

All steps of our analysis were implemented in annotated Python notebooks (https://github.com/paulfogel/A2SIGN). To perform NTF, we used the NMTF package, which can be downloaded using Python pip install.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Comparative assessment and novel strategy on methods for imputing proteomics data

Missing values are a major issue in quantitative proteomics analysis. While many methods have been developed for imputing missing values in high-throughput proteomics data, a comparative assessment of imputation accuracy remains inconclusive, mainly because mechanisms contributing to true missing values are complex and existing evaluation methodologies are imperfect. Moreover, few studies have provided an outlook of future methodological development. We first re-evaluate the performance of eight representative methods targeting three typical missing mechanisms. These methods are compared on both simulated and masked missing values embedded within real proteomics datasets, and performance is evaluated using three quantitative measures. We then introduce fused regularization matrix factorization, a low-rank global matrix factorization framework, capable of integrating local similarity derived from additional data types. We also explore a biologically-inspired latent variable modeling strategy—convex analysis of mixtures—for missing value imputation and present preliminary experimental results. While some winners emerged from our comparative assessment, the evaluation is intrinsically imperfect because performance is evaluated indirectly on artificial missing or masked values not authentic missing values. Nevertheless, we show that our fused regularization matrix factorization provides a novel incorporation of external and local information, and the exploratory implementation of convex analysis of mixtures presents a biologically plausible new approach.

Pancreas Whole Tissue Transcriptomics Highlights the Role of the Exocrine Pancreas in Patients With Recently Diagnosed Type 1 Diabetes

Although type 1 diabetes (T1D) is primarily a disease of the pancreatic beta-cells, understanding of the disease-associated alterations in the whole pancreas could be important for the improved treatment or the prevention of the disease. We have characterized the whole-pancreas gene expression of patients with recently diagnosed T1D from the Diabetes Virus Detection (DiViD) study and non-diabetic controls. Furthermore, another parallel dataset of the whole pancreas and an additional dataset from the laser-captured pancreatic islets of the DiViD patients and non-diabetic organ donors were analyzed together with the original dataset to confirm the results and to get further insights into the potential disease-associated differences between the exocrine and the endocrine pancreas. First, higher expression of the core acinar cell genes, encoding for digestive enzymes, was detected in the whole pancreas of the DiViD patients when compared to non-diabetic controls. Second, In the pancreatic islets, upregulation of immune and inflammation related genes was observed in the DiViD patients when compared to non-diabetic controls, in line with earlier publications, while an opposite trend was observed for several immune and inflammation related genes at the whole pancreas tissue level. Third, strong downregulation of the regenerating gene family (REG) genes, linked to pancreatic islet growth and regeneration, was observed in the exocrine acinar cell dominated whole-pancreas data of the DiViD patients when compared with the non-diabetic controls. Fourth, analysis of unique features in the transcriptomes of each DiViD patient compared with the other DiViD patients, revealed elevated expression of central antiviral immune response genes in the whole-pancreas samples, but not in the pancreatic islets, of one DiViD patient. This difference in the extent of antiviral gene expression suggests different statuses of infection in the pancreas at the time of sampling between the DiViD patients, who were all enterovirus VP1+ in the islets by immunohistochemistry based on earlier studies. The observed features, indicating differences in the function, status and interplay between the exocrine and the endocrine pancreas of recent onset T1D patients, highlight the importance of studying both compartments for better understanding of the molecular mechanisms of T1D.

Circulating RNAs in prostate cancer patients

Growing bodies of evidence have demonstrated that the identification of prostate cancer (PCa) biomarkers in the patients' blood and urine may remarkably improve PCa diagnosis and progression monitoring. Among diverse cancer-derived circulating materials, extracellular RNA molecules (exRNAs) represent a compelling component to investigate cancer-related alterations. Once outside the intracellular environment, exRNAs circulate in biofluids either in association with protein complexes or encapsulated inside extracellular vesicles (EVs). Notably, EV-associated RNAs (EV-RNAs) were used for the development of several assays (such as the FDA-approved Progensa Prostate Cancer Antigen 3 (PCA3 test) aiming at improving early PCa detection. EV-RNAs encompass a mixture of species, including small non-coding RNAs (e.g. miRNA and circRNA), lncRNAs and mRNAs. Several methods have been proposed to isolate EVs and relevant RNAs, and to perform RNA-Seq studies to identify potential cancer biomarkers. However, EVs in the circulation of a cancer patient include a multitude of diverse populations that are released by both cancer and normal cells from different tissues, thereby leading to a heterogeneous EV-RNA-associated transcriptional signal. Decrypting the complexity of such a composite signal is nowadays the major challenge faced in the identification of specific tumor-associated RNAs. Multiple deconvolution algorithms have been proposed so far to infer the enrichment of cancer-specific signals from gene expression data. However, novel strategies for EVs sorting and sequencing of RNA associated to single EVs populations will remarkably facilitate the identification of cancer-related molecules. Altogether, the studies summarized here demonstrate the high potential of using EV-RNA biomarkers in PCa and highlight the urgent need of improving technologies and computational approaches to characterize specific EVs populations and their relevant RNA cargo.

Deconvolution of Bulk Gene Expression Profiles with Single-Cell Transcriptomics to Develop a Cell Type Composition-Based Prognostic Model for Acute Myeloid Leukemia

Acute myeloid leukemia (AML) is one of the malignant hematologic cancers with rapid progress and poor prognosis. Most AML prognostic stratifications focused on genetic abnormalities. However, none of them was established based on the cell type compositions (CTCs) of peripheral blood or bone marrow aspirates from patients at diagnosis. Here we sought to develop a novel prognostic model for AML in adults based on the CTCs. First, we applied the CIBERSORT algorithm to estimate the CTCs for patients from two public datasets (GSE6891 and TCGA-LAML) using a custom gene expression signature reference constructed by an AML single-cell RNA sequencing dataset (GSE116256). Then, a CTC-based prognostic model was established using least absolute shrinkage and selection operator Cox regression, termed CTC score. The constructed prognostic model CTC score comprised 3 cell types, GMP-like, HSC-like, and T. Compared with the low-CTC-score group, the high-CTC-score group showed a 1.57-fold [95% confidence interval (CI), 1.23 to 2.00; p = 0.0002] and a 2.32-fold (95% CI, 1.53 to 3.51; p < 0.0001) higher overall mortality risk in the training set (GSE6891) and validation set (TCGA-LAML), respectively. When adjusting for age at diagnosis, cytogenetic risk, and karyotype, the CTC score remained statistically significant in both the training set [hazard ratio (HR) = 2.25; 95% CI, 1.20 to 4.24; p = 0.0119] and the validation set (HR = 7.97; 95% CI, 2.95 to 21.56; p < 0.0001]. We further compared the performance of the CTC score with two gene expression-based prognostic scores: the 17-gene leukemic stem cell score (LSC17 score) and the AML prognostic score (APS). It turned out that the CTC score achieved comparable performance at 1-, 2-, 3-, and 5-years timepoints and provided independent and additional prognostic information different from the LSC17 score and APS. In conclusion, the CTC score could serve as a powerful prognostic marker for AML and has great potential to assist clinicians to formulate individualized treatment plans.

Besca, a single-cell transcriptomics analysis toolkit to accelerate translational research

Single-cell RNA sequencing (scRNA-seq) revolutionized our understanding of disease biology. The promise it presents to also transform translational research requires highly standardized and robust software workflows. Here, we present the toolkit Besca, which streamlines scRNA-seq analyses and their use to deconvolute bulk RNA-seq data according to current best practices. Beyond a standard workflow covering quality control, filtering, and clustering, two complementary Besca modules, utilizing hierarchical cell signatures and supervised machine learning, automate cell annotation and provide harmonized nomenclatures. Subsequently, the gene expression profiles can be employed to estimate cell type proportions in bulk transcriptomics data. Using multiple, diverse scRNA-seq datasets, some stemming from highly heterogeneous tumor tissue, we show how Besca aids acceleration, interoperability, reusability and interpretability of scRNA-seq data analyses, meeting crucial demands in translational research and beyond.

Labour classified by cervical dilatation & fetal membrane rupture demonstrates differential impact on RNA-seq data for human myometrium tissues

High throughput sequencing has previously identified differentially expressed genes (DEGs) and enriched signalling networks in human myometrium for term (≥37 weeks) gestation labour, when defined as a singular state of activity at comparison to the non-labouring state. However, transcriptome changes that occur during transition from early to established labour (defined as ≤3 and >3 cm cervical dilatation, respectively) and potentially altered by fetal membrane rupture (ROM), when adapting from onset to completion of childbirth, remained to be defined. In the present study, we assessed whether differences for these two clinically observable factors of labour are associated with different myometrial transcriptome profiles. Analysis of our tissue (‘bulk’) RNA-seq data (NCBI Gene Expression Omnibus: GSE80172) with classification of labour into four groups, each compared to the same non-labour group, identified more DEGs for early than established labour; ROM was the strongest up-regulator of DEGs. We propose that lower DEGs frequency for early labour and/or ROM negative myometrium was attributed to bulk RNA-seq limitations associated with tissue heterogeneity, as well as the possibility that processes other than gene transcription are of more importance at labour onset. Integrative analysis with future data from additional samples, which have at least equivalent refined clinical classification for labour status, and alternative omics approaches will help to explain what truly contributes to transcriptomic changes that are critical for labour onset. Lastly, we identified five DEGs common to all labour groupings; two of which (AREG and PER3) were validated by qPCR and not differentially expressed in placenta and choriodecidua.

Single-Cell Genomics: Enabling the Functional Elucidation of Infectious Diseases in Multi-Cell Genomes

Since the time when detection of gene expression in single cells by microarrays to the Next Generation Sequencing (NGS) enabled Single Cell Genomics (SCG), it has played a pivotal role to understand and elucidate the functional role of cellular heterogeneity. Along this journey to becoming a key player in the capture of the individuality of cells, SCG overcame many milestones, including scale, speed, sensitivity and sample costs (4S). There have been many important experimental and computational innovations in the efficient analysis and interpretation of SCG data. The increasing role of AI in SCG data analysis has further enhanced its applicability in building models for clinical intervention. Furthermore, SCG has been instrumental in the delineation of the role of cellular heterogeneity in specific diseases, including cancer and infectious diseases. The understanding of the role of differential immune responses in driving coronavirus disease-2019 (COVID-19) disease severity and clinical outcomes has been greatly aided by SCG. With many variants of concern (VOC) in sight, it would be of great importance to further understand the immune response specificity vis-a-vis the immune cell repertoire, the identification of novel cell types, and antibody response. Given the potential of SCG to play an integral part in the multi-omics approach to the study of the host-pathogen interaction and its outcomes, our review attempts to highlight its strengths, its implications for infectious disease biology, and its current limitations. We conclude that the application of SCG would be a critical step towards future pandemic preparedness.

Fibromine is a multi-omics database and mining tool for target discovery in pulmonary fibrosis

Idiopathic pulmonary fibrosis is a lethal lung fibroproliferative disease with limited therapeutic options. Differential expression profiling of affected sites has been instrumental for involved pathogenetic mechanisms dissection and therapeutic targets discovery. However, there have been limited efforts to comparatively analyse/mine the numerous related publicly available datasets, to fully exploit their potential on the validation/creation of novel research hypotheses. In this context and towards that goal, we present Fibromine, an integrated database and exploration environment comprising of consistently re-analysed, manually curated transcriptomic and proteomic pulmonary fibrosis datasets covering a wide range of experimental designs in both patients and animal models. Fibromine can be accessed via an R Shiny application (http://www.fibromine.com/Fibromine) which offers dynamic data exploration and real-time integration functionalities. Moreover, we introduce a novel benchmarking system based on transcriptomic datasets underlying characteristics, resulting to dataset accreditation aiming to aid the user on dataset selection. Cell specificity of gene expression can be visualised and/or explored in several scRNA-seq datasets, in an effort to link legacy data with this cutting-edge methodology and paving the way to their integration. Several use case examples are presented, that, importantly, can be reproduced on-the-fly by a non-specialist user, the primary target and potential user of this endeavour.

Identification of cell-type-specific marker genes from co-expression patterns in tissue samples

MOTIVATION

Marker genes, defined as genes that are expressed primarily in a single-cell type, can be identified from the single-cell transcriptome; however, such data are not always available for the many uses of marker genes, such as deconvolution of bulk tissue. Marker genes for a cell type, however, are highly correlated in bulk data, because their expression levels depend primarily on the proportion of that cell type in the samples. Therefore, when many tissue samples are analyzed, it is possible to identify these marker genes from the correlation pattern.

RESULTS

To capitalize on this pattern, we develop a new algorithm to detect marker genes by combining published information about likely marker genes with bulk transcriptome data in the form of a semi-supervised algorithm. The algorithm then exploits the correlation structure of the bulk data to refine the published marker genes by adding or removing genes from the list.

AVAILABILITY AND IMPLEMENTATION

We implement this method as an R package markerpen, hosted on CRAN (https://CRAN.R-project.org/package=markerpen).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Machine learning applications for therapeutic tasks with genomics data

Thanks to the increasing availability of genomics and other biomedical data, many machine learning algorithms have been proposed for a wide range of therapeutic discovery and development tasks. In this survey, we review the literature on machine learning applications for genomics through the lens of therapeutic development. We investigate the interplay among genomics, compounds, proteins, electronic health records, cellular images, and clinical texts. We identify 22 machine learning in genomics applications that span the whole therapeutics pipeline, from discovering novel targets, personalizing medicine, developing gene-editing tools, all the way to facilitating clinical trials and post-market studies. We also pinpoint seven key challenges in this field with potentials for expansion and impact. This survey examines recent research at the intersection of machine learning, genomics, and therapeutic development.

DECONbench: a benchmarking platform dedicated to deconvolution methods for tumor heterogeneity quantification

BACKGROUND

Quantification of tumor heterogeneity is essential to better understand cancer progression and to adapt therapeutic treatments to patient specificities. Bioinformatic tools to assess the different cell populations from single-omic datasets as bulk transcriptome or methylome samples have been recently developed, including reference-based and reference-free methods. Improved methods using multi-omic datasets are yet to be developed in the future and the community would need systematic tools to perform a comparative evaluation of these algorithms on controlled data.

RESULTS

We present DECONbench, a standardized unbiased benchmarking resource, applied to the evaluation of computational methods quantifying cell-type heterogeneity in cancer. DECONbench includes gold standard simulated benchmark datasets, consisting of transcriptome and methylome profiles mimicking pancreatic adenocarcinoma molecular heterogeneity, and a set of baseline deconvolution methods (reference-free algorithms inferring cell-type proportions). DECONbench performs a systematic performance evaluation of each new methodological contribution and provides the possibility to publicly share source code and scoring.

CONCLUSION

DECONbench allows continuous submission of new methods in a user-friendly fashion, each novel contribution being automatically compared to the reference baseline methods, which enables crowdsourced benchmarking. DECONbench is designed to serve as a reference platform for the benchmarking of deconvolution methods in the evaluation of cancer heterogeneity. We believe it will contribute to leverage the benchmarking practices in the biomedical and life science communities. DECONbench is hosted on the open source Codalab competition platform. It is freely available at: https://competitions.codalab.org/competitions/27453 .

Integrating single-cell and spatial transcriptomics to elucidate intercellular tissue dynamics

Single-cell RNA sequencing (scRNA-seq) identifies cell subpopulations within tissue but does not capture their spatial distribution nor reveal local networks of intercellular communication acting in situ. A suite of recently developed techniques that localize RNA within tissue, including multiplexed in situ hybridization and in situ sequencing (here defined as high-plex RNA imaging) and spatial barcoding, can help address this issue. However, no method currently provides as complete a scope of the transcriptome as does scRNA-seq, underscoring the need for approaches to integrate single-cell and spatial data. Here, we review efforts to integrate scRNA-seq with spatial transcriptomics, including emerging integrative computational methods, and propose ways to effectively combine current methodologies.

SCISSOR™: a single-cell inferred site-specific omics resource for tumor microenvironment association study

Tumor tissues are heterogeneous with different cell types in tumor microenvironment, which play an important role in tumorigenesis and tumor progression. Several computational algorithms and tools have been developed to infer the cell composition from bulk transcriptome profiles. However, they ignore the tissue specificity and thus a new resource for tissue-specific cell transcriptomic reference is needed for inferring cell composition in tumor microenvironment and exploring their association with clinical outcomes and tumor omics. In this study, we developed SCISSOR™ (https://thecailab.com/scissor/), an online open resource to fulfill that demand by integrating five orthogonal omics data of >6031 large-scale bulk samples, patient clinical outcomes and 451 917 high-granularity tissue-specific single-cell transcriptomic profiles of 16 cancer types. SCISSOR™ provides five major analysis modules that enable flexible modeling with adjustable parameters and dynamic visualization approaches. SCISSOR™ is valuable as a new resource for promoting tumor heterogeneity and tumor–tumor microenvironment cell interaction research, by delineating cells in the tissue-specific tumor microenvironment and characterizing their associations with tumor omics and clinical outcomes.

ImmuCellDB: An Indicative Database of Immune Cell Composition From Different Tissues and Disease Conditions in Mouse and Human

Immune cell composition is highly divergent across different tissues and diseases. A comprehensive resource of tissue immune cells across different conditions in mouse and human will thus provide great understanding of the immune microenvironment of many diseases. Recently, computational methods for estimating immune cell abundance from tissue transcriptome data have been developed and are now widely used. Using these computational tools, large-scale estimation of immune cell composition across tissues and conditions should be possible using gene expression data collected from public databases. In total, 266 tissue types and 706 disease types in humans, as well as 143 tissue types and 61 disease types, and 206 genotypes in mouse had been included in a database we have named ImmuCellDB (http://wap-lab.org:3200/ImmuCellDB/). In ImmuCellDB, users can search and browse immune cell proportions based on tissues, disease or genotype in mouse or humans. Additionally, the variation and correlation of immune cell abundance and gene expression level between different conditions can be compared and viewed in this database. We believe that ImmuCellDB provides not only an indicative view of tissue-dependent or disease-dependent immune cell profiles, but also represents an easy way to pre-determine immune cell abundance and gene expression profiles for specific situations.

Molecular characterization of projection neuron subtypes in the mouse olfactory bulb

Projection neurons (PNs) in the mammalian olfactory bulb (OB) receive input from the nose and project to diverse cortical and subcortical areas. Morphological and physiological studies have highlighted functional heterogeneity, yet no molecular markers have been described that delineate PN subtypes. Here, we used viral injections into olfactory cortex and fluorescent nucleus sorting to enrich PNs for high-throughput single nucleus and bulk RNA deep sequencing. Transcriptome analysis and RNA in situ hybridization identified distinct mitral and tufted cell populations with characteristic transcription factor network topology, cell adhesion, and excitability-related gene expression. Finally, we describe a new computational approach for integrating bulk and snRNA-seq data and provide evidence that different mitral cell populations preferentially project to different target regions. Together, we have identified potential molecular and gene regulatory mechanisms underlying PN diversity and provide new molecular entry points into studying the diverse functional roles of mitral and tufted cell subtypes.

A likelihood-based deconvolution of bulk gene expression data using single-cell references

Direct comparison of bulk gene expression profiles is complicated by distinct cell type mixtures in each sample which obscure whether observed differences are actually due to changes in expression levels themselves or simply due to differing cell type compositions. Single-cell technology has made it possible to measure gene expression in individual cells, achieving higher resolution at the expense of increased noise. If carefully incorporated, such single-cell data can be used to deconvolve bulk samples to yield accurate estimates of the true cell type proportions, thus enabling one to disentangle the effects of differential expression and cell type mixtures. Here, we propose a generative model and a likelihood-based inference method that uses asymptotic statistical theory and a novel optimization procedure to perform deconvolution of bulk RNA-seq data to produce accurate cell type proportion estimates. We demonstrate the effectiveness of our method, called RNA-Sieve, across a diverse array of scenarios involving real data and discuss extensions made uniquely possible by our probabilistic framework, including a demonstration of well-calibrated confidence intervals.

Transfer transcriptomic signatures for infectious diseases

Human responses to infection include transcriptional changes shared across diverse pathogens. To capture these common patterns, we establish the concept of, and the method for, the identification of “transfer signatures”: sets of genes defining human immunophenotypes. We demonstrate the usefulness of transfer signatures in two use cases: the progression of latent to active tuberculosis and the severity of viral respiratory infections.

The modulation of the transcriptome is among the earliest responses to infection. However, defining the transcriptomic signatures of disease is challenging because logistic, technical, and cost factors limit the size and representativeness of samples in clinical studies. These limitations lead to a poor performance of signatures when applied to new datasets. Although the study focuses on infection, the central hypothesis of the work is the generalization of sets of signatures across diseases. We use a machine learning approach to identify common elements in datasets and then test empirically whether they are informative about a second dataset from a disease or process distinct from the original dataset. We identify sets of genes, which we name transfer signatures, that are predictive across diverse datasets and/or species (e.g., rhesus to humans). We demonstrate the usefulness of transfer signatures in two use cases: the progression of latent to active tuberculosis and the severity of COVID-19 and influenza A H1N1 infection. This indicates that transfer signatures can be deployed in settings that lack disease-specific biomarkers. The broad significance of our work lies in the concept that a small set of archetypal human immunophenotypes, captured by transfer signatures, can explain a larger set of responses to diverse diseases.

Applications of single-cell and bulk RNA sequencing in onco-immunology

The rising interest for precise characterization of the tumour immune contexture has recently brought forward the high potential of RNA sequencing (RNA-seq) in identifying molecular mechanisms engaged in the response to immunotherapy. In this review, we provide an overview of the major principles of single-cell and conventional (bulk) RNA-seq applied to onco-immunology. We describe standard preprocessing and statistical analyses of data obtained from such techniques and highlight some computational challenges relative to the sequencing of individual cells. We notably provide examples of gene expression analyses such as differential expression analysis, dimensionality reduction, clustering and enrichment analysis. Additionally, we used public data sets to exemplify how deconvolution algorithms can identify and quantify multiple immune subpopulations from either bulk or single-cell RNA-seq. We give examples of machine and deep learning models used to predict patient outcomes and treatment effect from high-dimensional data. Finally, we balance the strengths and weaknesses of single-cell and bulk RNA-seq regarding their applications in the clinic.

A benchmark for RNA-seq deconvolution analysis under dynamic testing environments

BACKGROUND

Deconvolution analyses have been widely used to track compositional alterations of cell types in gene expression data. Although a large number of novel methods have been developed, due to a lack of understanding of the effects of modeling assumptions and tuning parameters, it is challenging for researchers to select an optimal deconvolution method suitable for the targeted biological conditions.

RESULTS

To systematically reveal the pitfalls and challenges of deconvolution analyses, we investigate the impact of several technical and biological factors including simulation model, quantification unit, component number, weight matrix, and unknown content by constructing three benchmarking frameworks. These frameworks cover comparative analysis of 11 popular deconvolution methods under 1766 conditions.

CONCLUSIONS

We provide new insights to researchers for future application, standardization, and development of deconvolution tools on RNA-seq data.

Comparison of genome-wide gene expression profiling by RNA Sequencing versus microarray in bronchial biopsies of COPD patients before and after inhaled corticosteroid treatment: does it provide new insights?

More DEGs are detected by RNA-Seq than microarrays in COPD lung biopsies and are associated with immunological pathways. Performing bulk tissue cell-type deconvolution in microarray lung samples, using the SVR method, reflects RNA-Seq results. https://bit.ly/2N8sY3s.

Advances in bulk and single-cell multi-omics approaches for systems biology and precision medicine

Multi-omics allows the systematic understanding of the information flow across different omics layers, while single omics can mainly reflect one aspect of the biological system. The advancement of bulk and single-cell sequencing technologies and related computational methods for multi-omics largely facilitated the development of system biology and precision medicine. Single-cell approaches have the advantage of dissecting cellular dynamics and heterogeneity, whereas traditional bulk technologies are limited to individual/population-level investigation. In this review, we first summarize the technologies for producing bulk and single-cell multi-omics data. Then, we survey the computational approaches for integrative analysis of bulk and single-cell multimodal data, respectively. Moreover, the databases and data storage for multi-omics, as well as the tools for visualizing multimodal data are summarized. We also outline the integration between bulk and single-cell data, and discuss the applications of multi-omics in precision medicine. Finally, we present the challenges and perspectives for multi-omics development.

Nonlinear ridge regression improves cell-type-specific differential expression analysis

BACKGROUND

Epigenome-wide association studies (EWAS) and differential gene expression analyses are generally performed on tissue samples, which consist of multiple cell types. Cell-type-specific effects of a trait, such as disease, on the omics expression are of interest but difficult or costly to measure experimentally. By measuring omics data for the bulk tissue, cell type composition of a sample can be inferred statistically. Subsequently, cell-type-specific effects are estimated by linear regression that includes terms representing the interaction between the cell type proportions and the trait. This approach involves two issues, scaling and multicollinearity.

RESULTS

First, although cell composition is analyzed in linear scale, differential methylation/expression is analyzed suitably in the logit/log scale. To simultaneously analyze two scales, we applied nonlinear regression. Second, we show that the interaction terms are highly collinear, which is obstructive to ordinary regression. To cope with the multicollinearity, we applied ridge regularization. In simulated data, nonlinear ridge regression attained well-balanced sensitivity, specificity and precision. Marginal model attained the lowest precision and highest sensitivity and was the only algorithm to detect weak signal in real data.

CONCLUSION

Nonlinear ridge regression performed cell-type-specific association test on bulk omics data with well-balanced performance. The omicwas package for R implements nonlinear ridge regression for cell-type-specific EWAS, differential gene expression and QTL analyses. The software is freely available from https://github.com/fumi-github/omicwas.

SCDC: bulk gene expression deconvolution by multiple single-cell RNA sequencing references

Recent advances in single-cell RNA sequencing (scRNA-seq) enable characterization of transcriptomic profiles with single-cell resolution and circumvent averaging artifacts associated with traditional bulk RNA sequencing (RNA-seq) data. Here, we propose SCDC, a deconvolution method for bulk RNA-seq that leverages cell-type specific gene expression profiles from multiple scRNA-seq reference datasets. SCDC adopts an ENSEMBLE method to integrate deconvolution results from different scRNA-seq datasets that are produced in different laboratories and at different times, implicitly addressing the problem of batch-effect confounding. SCDC is benchmarked against existing methods using both in silico generated pseudo-bulk samples and experimentally mixed cell lines, whose known cell-type compositions serve as ground truths. We show that SCDC outperforms existing methods with improved accuracy of cell-type decomposition under both settings. To illustrate how the ENSEMBLE framework performs in complex tissues under different scenarios, we further apply our method to a human pancreatic islet dataset and a mouse mammary gland dataset. SCDC returns results that are more consistent with experimental designs and that reproduce more significant associations between cell-type proportions and measured phenotypes.

Computational deconvolution to estimate cell type-specific gene expression from bulk data

Computational deconvolution is a time and cost-efficient approach to obtain cell type-specific information from bulk gene expression of heterogeneous tissues like blood. Deconvolution can aim to either estimate cell type proportions or abundances in samples, or estimate how strongly each present cell type expresses different genes, or both tasks simultaneously. Among the two separate goals, the estimation of cell type proportions/abundances is widely studied, but less attention has been paid on defining the cell type-specific expression profiles. Here, we address this gap by introducing a novel method Rodeo and empirically evaluating it and the other available tools from multiple perspectives utilizing diverse datasets.

Progress and challenge for computational quantification of tissue immune cells

Tissue immune cells have long been recognized as important regulators for the maintenance of balance in the body system. Quantification of the abundance of different immune cells will provide enhanced understanding of the correlation between immune cells and normal or abnormal situations. Currently, computational methods to predict tissue immune cell compositions from bulk transcriptomes have been largely developed. Therefore, summarizing the advantages and disadvantages is appropriate. In addition, an examination of the challenges and possible solutions for these computational models will assist the development of this field. The common hypothesis of these models is that the expression of signature genes for immune cell types might represent the proportion of immune cells that contribute to the tissue transcriptome. In general, we grouped all reported tools into three groups, including reference-free, reference-based scoring and reference-based deconvolution methods. In this review, a summary of all the currently reported computational immune cell quantification tools and their applications, limitations, and perspectives are presented. Furthermore, some critical problems are found that have limited the performance and application of these models, including inadequate immune cell type, the collinearity problem, the impact of the tissue environment on the immune cell expression level, and the deficiency of standard datasets for model validation. To address these issues, tissue specific training datasets that include all known immune cells, a hierarchical computational framework, and benchmark datasets including both tissue expression profiles and the abundances of all the immune cells are proposed to further promote the development of this field.

Systems Biology Analysis for Ewing Sarcoma

Ewing sarcoma (EwS) is a highly aggressive pediatric bone cancer that is defined by a somatic fusion between the EWSR1 gene and an ETS family member, most frequently the FLI1 gene, leading to expression of a chimeric transcription factor EWSR1-FLI1. Otherwise, EwS is one of the most genetically stable cancers. The situation when the major cancer driver is well known looks like a unique opportunity for applying the systems biology approach in order to understand the EwS mechanisms as well as to uncover some general mechanistic principles of carcinogenesis. A number of studies have been performed revealing the direct and indirect effects of EWSR1-FLI1 on multiple aspects of cellular life. Nevertheless, the emerging picture of the oncogene action appears to be highly complex and systemic, with multiple reciprocal influences between the immediate consequences of the driver mutation and intracellular and intercellular molecular mechanisms, including regulation of transcription, epigenome, and tumoral microenvironment. In this chapter, we present an overview of existing molecular profiling resources available for EwS tumors and cell lines and provide an online comprehensive catalogue of publicly available omics and other datasets. We further highlight the systems biology studies of EwS, involving mathematical modeling of networks and integration of molecular data. We conclude that despite the seeming simplicity, a lot has yet to be understood on the systems-wide mechanisms connecting the driver mutation and the major cellular phenotypes of this pediatric cancer. Overall, this chapter can serve as a guide for a systems biology researcher to start working on EwS.