Pathway-level information extractor (PLIER) for gene expression data

Signature reversion of three disease-associated gene signatures prioritizes cancer drug repurposing candidates

Drug repurposing is promising because approving a drug for a new indication requires fewer resources than approving a new drug. Signature reversion detects drug perturbations most inversely related to the disease-associated gene signature to identify drugs that may reverse that signature. We assessed the performance and biological relevance of three approaches for constructing disease-associated gene signatures (i.e., limma, DESeq2, and MultiPLIER) and prioritized the resulting drug repurposing candidates for four low-survival human cancers. Our results were enriched for candidates that had been used in clinical trials or performed well in the PRISM drug screen. Additionally, we found that pamidronate and nimodipine, drugs predicted to be efficacious against the brain tumor glioblastoma (GBM), inhibited the growth of a GBM cell line and cells isolated from a patient-derived xenograft (PDX). Our results demonstrate that by applying multiple disease-associated gene signature methods, we prioritized several drug repurposing candidates for low-survival cancers.

Molecular adaptations in response to exercise training are associated with tissue-specific transcriptomic and epigenomic signatures

Regular exercise has many physical and brain health benefits, yet the molecular mechanisms mediating exercise effects across tissues remain poorly understood. Here we analyzed 400 high-quality DNA methylation, ATAC-seq, and RNA-seq datasets from eight tissues from control and endurance exercise-trained (EET) rats. Integration of baseline datasets mapped the gene location dependence of epigenetic control features and identified differing regulatory landscapes in each tissue. The transcriptional responses to 8 weeks of EET showed little overlap across tissues and predominantly comprised tissue-type enriched genes. We identified sex differences in the transcriptomic and epigenomic changes induced by EET. However, the sex-biased gene responses were linked to shared signaling pathways. We found that many G protein-coupled receptor-encoding genes are regulated by EET, suggesting a role for these receptors in mediating the molecular adaptations to training across tissues. Our findings provide new insights into the mechanisms underlying EET-induced health benefits across organs.

Pathway Activation Analysis for Pan-Cancer Personalized Characterization Based on Riemannian Manifold

The pathogenesis of carcinoma is believed to come from the combined effect of polygenic variation, and the initiation and progression of malignant tumors are closely related to the dysregulation of biological pathways. Quantifying the alteration in pathway activation and identifying coordinated patterns of pathway dysfunction are the imperative part of understanding the malignancy process and distinguishing different tumor stages or clinical outcomes of individual patients. In this study, we have conducted in silico pathway activation analysis using Riemannian manifold (RiePath) toward pan-cancer personalized characterization, which is the first attempt to apply the Riemannian manifold theory to measure the extent of pathway dysregulation in individual patient on the tangent space of the Riemannian manifold. RiePath effectively integrates pathway and gene expression information, not only generating a relatively low-dimensional and biologically relevant representation, but also identifying a robust panel of biologically meaningful pathway signatures as biomarkers. The pan-cancer analysis across 16 cancer types reveals the capability of RiePath to evaluate pathway activation accurately and identify clinical outcome-related pathways. We believe that RiePath has the potential to provide new prospects in understanding the molecular mechanisms of complex diseases and may find broader applications in predicting biomarkers for other intricate diseases.

Regional and bilateral MRI and gene signatures in facioscapulohumeral dystrophy: implications for clinical trial design and mechanisms of disease progression

Identifying the aberrant expression of DUX4 in skeletal muscle as the cause of facioscapulohumeral dystrophy (FSHD) has led to rational therapeutic development and clinical trials. Several studies support the use of MRI characteristics and the expression of DUX4-regulated genes in muscle biopsies as biomarkers of FSHD disease activity and progression. We performed lower-extremity MRI and muscle biopsies in the mid-portion of the tibialis anterior (TA) muscles bilaterally in FSHD subjects and validated our prior reports of the strong association between MRI characteristics and expression of genes regulated by DUX4 and other gene categories associated with FSHD disease activity. We further show that measurements of normalized fat content in the entire TA muscle strongly predict molecular signatures in the mid-portion of the TA, indicating that regional biopsies can accurately measure progression in the whole muscle and providing a strong basis for inclusion of MRI and molecular biomarkers in clinical trial design. An unanticipated finding was the strong correlations of molecular signatures in the bilateral comparisons, including markers of B-cells and other immune cell populations, suggesting that a systemic immune cell infiltration of skeletal muscle might have a role in disease progression.

BuDDI: BulkDeconvolution withDomainInvariance to predict cell-type-specific perturbations from bulk

While single-cell experiments provide deep cellular resolution within a single sample, some single-cell experiments are inherently more challenging than bulk experiments due to dissociation difficulties, cost, or limited tissue availability. This creates a situation where we have deep cellular profiles of one sample or condition, and bulk profiles across multiple samples and conditions. To bridge this gap, we propose BuDDI (BUlk Deconvolution with Domain Invariance). BuDDI utilizes domain adaptation techniques to effectively integrate available corpora of case-control bulk and reference scRNA-seq observations to infer cell-type-specific perturbation effects. BuDDI achieves this by learning independent latent spaces within a single variational autoencoder (VAE) encompassing at least four sources of variability: 1) cell type proportion, 2) perturbation effect, 3) structured experimental variability, and 4) remaining variability. Since each latent space is encouraged to be independent, we simulate perturbation responses by independently composing each latent space to simulate cell-type-specific perturbation responses. We evaluated BuDDI's performance on simulated and real data with experimental designs of increasing complexity. We first validated that BuDDI could learn domain invariant latent spaces on data with matched samples across each source of variability. Then we validated that BuDDI could accurately predict cell-type-specific perturbation response when no single-cell perturbed profiles were used during training; instead, only bulk samples had both perturbed and non-perturbed observations. Finally, we validated BuDDI on predicting sex-specific differences, an experimental design where it is not possible to have matched samples. In each experiment, BuDDI outperformed all other comparative methods and baselines. As more reference atlases are completed, BuDDI provides a path to combine these resources with bulk-profiled treatment or disease signatures to study perturbations, sex differences, or other factors at single-cell resolution.

A transfer learning framework to elucidate the clinical relevance of altered proximal tubule cell states in kidney disease

The application of single-cell technologies in clinical nephrology remains elusive. We generated an atlas of transcriptionally defined cell types and cell states of human kidney disease by integrating single-cell signatures reported in the literature with newly generated signatures obtained from 5 patients with acute kidney injury. We used this information to develop kidney-specific cell-level information ExtractoR (K-CLIER), a transfer learning approach specifically tailored to evaluate the role of cell types/states on bulk RNAseq data. We validated the K-CLIER as a reliable computational framework to obtain a dimensionality reduction and to link clinical data with single-cell signatures. By applying K-CLIER on cohorts of patients with different kidney diseases, we identified the most relevant cell types associated with fibrosis and disease progression. This analysis highlighted the central role of altered proximal tubule cells in chronic kidney disease. Our study introduces a new strategy to exploit the power of single-cell technologies toward clinical applications.

DiseaseNet: a transfer learning approach to noncommunicable disease classification

As noncommunicable diseases (NCDs) pose a significant global health burden, identifying effective diagnostic and predictive markers for these diseases is of paramount importance. Epigenetic modifications, such as DNA methylation, have emerged as potential indicators for NCDs. These have previously been exploited in other contexts within the framework of neural network models that capture complex relationships within the data. Applications of neural networks have led to significant breakthroughs in various biological or biomedical fields but these have not yet been effectively applied to NCD modeling. This is, in part, due to limited datasets that are not amenable to building of robust neural network models. In this work, we leveraged a neural network trained on one class of NCDs, cancer, as the basis for a transfer learning approach to non-cancer NCD modeling. Our results demonstrate promising performance of the model in predicting three NCDs, namely, arthritis, asthma, and schizophrenia, for the respective blood samples, with an overall accuracy (f-measure) of 94.5%. Furthermore, a concept based explanation method called Testing with Concept Activation Vectors (TCAV) was used to investigate the importance of the sample sources and understand how future training datasets for multiple NCD models may be improved. Our findings highlight the effectiveness of transfer learning in developing accurate diagnostic and predictive models for NCDs.

Advances in cancer DNA methylation analysis with methPLIER: use of non-negative matrix factorization and knowledge-based constraints to enhance biological interpretability

DNA methylation is an epigenetic modification that results in dynamic changes during ontogenesis and cell differentiation. DNA methylation patterns regulate gene expression and have been widely researched. While tools for DNA methylation analysis have been developed, most of them have focused on intergroup comparative analysis within a dataset; therefore, it is difficult to conduct cross-dataset studies, such as rare disease studies or cross-institutional studies. This study describes a novel method for DNA methylation analysis, namely, methPLIER, which enables interdataset comparative analyses. methPLIER combines Pathway Level Information Extractor (PLIER), which is a non-negative matrix factorization (NMF) method, with regularization by a knowledge matrix and transfer learning. methPLIER can be used to perform intersample and interdataset comparative analysis based on latent feature matrices, which are obtained via matrix factorization of large-scale data, and factor-loading matrices, which are obtained through matrix factorization of the data to be analyzed. We used methPLIER to analyze a lung cancer dataset and confirmed that the data decomposition reflected sample characteristics for recurrence-free survival. Moreover, methPLIER can analyze data obtained via different preprocessing methods, thereby reducing distributional bias among datasets due to preprocessing. Furthermore, methPLIER can be employed for comparative analyses of methylation data obtained from different platforms, thereby reducing bias in data distribution due to platform differences. methPLIER is expected to facilitate cross-sectional DNA methylation data analysis and enhance DNA methylation data resources.

Fast and slow muscle fiber transcriptome dynamics with lifelong endurance exercise

We investigated fast and slow muscle fiber transcriptome exercise dynamics among three groups of men: lifelong exercisers (LLE, n = 8, 74 ± 1 yr), old healthy nonexercisers (OH, n = 9, 75 ± 1 yr), and young exercisers (YE, n = 8, 25 ± 1 yr). On average, LLE had exercised ∼4 day·wk-1 for ∼8 h·wk-1 over 53 ± 2 years. Muscle biopsies were obtained pre- and 4 h postresistance exercise (3 × 10 knee extensions at 70% 1-RM). Fast and slow fiber size and function were assessed preexercise with fast and slow RNA-seq profiles examined pre- and postexercise. LLE fast fiber size was similar to OH, which was ∼30% smaller than YE (P < 0.05) with contractile function variables among groups, resulting in lower power in LLE (P < 0.05). LLE slow fibers were ∼30% larger and more powerful compared with YE and OH (P < 0.05). At the transcriptome level, fast fibers were more responsive to resistance exercise compared with slow fibers among all three cohorts (P < 0.05). Exercise induced a comprehensive biological response in fast fibers (P < 0.05) including transcription, signaling, skeletal muscle cell differentiation, and metabolism with vast differences among the groups. Fast fibers from YE exhibited a growth and metabolic signature, with LLE being primarily metabolic, and OH showing a strong stress-related response. In slow fibers, only LLE exhibited a biological response to exercise (P < 0.05), which was related to ketone and lipid metabolism. The divergent exercise transcriptome signatures provide novel insight into the molecular regulation in fast and slow fibers with age and exercise and suggest that the ∼5% weekly exercise time commitment of the lifelong exercisers provided a powerful investment for fast and slow muscle fiber metabolic health at the molecular level.NEW & NOTEWORTHY This study provides the first insights into fast and slow muscle fiber transcriptome dynamics with lifelong endurance exercise. The fast fibers were more responsive to exercise with divergent transcriptome signatures among young exercisers (growth and metabolic), lifelong exercisers (metabolic), and old healthy nonexercisers (stress). Only lifelong exercisers had a biological response in slow fibers (metabolic). These data provide novel insights into fast and slow muscle fiber health at the molecular level with age and exercise.

FSHD muscle shows perturbation in fibroadipogenic progenitor cells, mitochondrial function and alternative splicing independently of inflammation

Facioscapulohumeral muscular dystrophy (FSHD) is a prevalent, incurable myopathy. FSHD is highly heterogeneous, with patients following a variety of clinical trajectories, complicating clinical trials. Skeletal muscle in FSHD undergoes fibrosis and fatty replacement that can be accelerated by inflammation, adding to heterogeneity. Well controlled molecular studies are thus essential to both categorize FSHD patients into distinct subtypes and understand pathomechanisms. Here, we further analyzed RNA-sequencing data from 24 FSHD patients, each of whom donated a biopsy from both a non-inflamed (TIRM-) and inflamed (TIRM+) muscle, and 15 FSHD patients who donated peripheral blood mononucleated cells (PBMCs), alongside non-affected control individuals. Differential gene expression analysis identified suppression of mitochondrial biogenesis and up-regulation of fibroadipogenic progenitor (FAP) gene expression in FSHD muscle, which was particularly marked on inflamed samples. PBMCs demonstrated suppression of antigen presentation in FSHD. Gene expression deconvolution revealed FAP expansion as a consistent feature of FSHD muscle, via meta-analysis of 7 independent transcriptomic datasets. Clustering of muscle biopsies separated patients in an unbiased manner into clinically mild and severe subtypes, independently of known disease modifiers (age, sex, D4Z4 repeat length). Lastly, the first genome-wide analysis of alternative splicing in FSHD muscle revealed perturbation of autophagy, BMP2 and HMGB1 signalling. Overall, our findings reveal molecular subtypes of FSHD with clinical relevance and identify novel pathomechanisms for this highly heterogeneous condition.

Exploring the effects of experimental parameters and data modeling approaches on in vitro transcriptomic point-of-departure estimates

Multiple new approach methods (NAMs) are being developed to rapidly screen large numbers of chemicals to aid in hazard evaluation and risk assessments. High-throughput transcriptomics (HTTr) in human cell lines has been proposed as a first-tier screening approach for determining the types of bioactivity a chemical can cause (activation of specific targets vs. generalized cell stress) and for calculating transcriptional points of departure (tPODs) based on changes in gene expression. In the present study, we examine a range of computational methods to calculate tPODs from HTTr data, using six data sets in which MCF7 cells cultured in two different media formulations were treated with a panel of 44 chemicals for 3 different exposure durations (6, 12, 24 hr). The tPOD calculation methods use data at the level of individual genes and gene set signatures, and compare data processed using the ToxCast Pipeline 2 (tcplfit2), BMDExpress and PLIER (Pathway Level Information ExtractoR). Methods were evaluated by comparing to in vitro PODs from a validated set of high-throughput screening (HTS) assays for a set of estrogenic compounds. Key findings include: (1) for a given chemical and set of experimental conditions, tPODs calculated by different methods can vary by several orders of magnitude; (2) tPODs are at least as sensitive to computational methods as to experimental conditions; (3) in comparison to an external reference set of PODs, some methods give generally higher values, principally PLIER and BMDExpress; and (4) the tPODs from HTTr in this one cell type are mostly higher than the overall PODs from a broad battery of targeted in vitro ToxCast assays, reflecting the need to test chemicals in multiple cell types and readout technologies for in vitro hazard screening.

A new Bayesian factor analysis method improves detection of genes and biological processes affected by perturbations in single-cell CRISPR screening

Clustered regularly interspaced short palindromic repeats (CRISPR) screening coupled with single-cell RNA sequencing has emerged as a powerful tool to characterize the effects of genetic perturbations on the whole transcriptome at a single-cell level. However, due to its sparsity and complex structure, analysis of single-cell CRISPR screening data is challenging. In particular, standard differential expression analysis methods are often underpowered to detect genes affected by CRISPR perturbations. We developed a statistical method for such data, called guided sparse factor analysis (GSFA). GSFA infers latent factors that represent coregulated genes or gene modules; by borrowing information from these factors, it infers the effects of genetic perturbations on individual genes. We demonstrated through extensive simulation studies that GSFA detects perturbation effects with much higher power than state-of-the-art methods. Using single-cell CRISPR data from human CD8+ T cells and neural progenitor cells, we showed that GSFA identified biologically relevant gene modules and specific genes affected by CRISPR perturbations, many of which were missed by existing methods, providing new insights into the functions of genes involved in T cell activation and neurodevelopment.

Integrated single-cell multiome analysis reveals muscle fiber-type gene regulatory circuitry modulated by endurance exercise

Endurance exercise is an important health modifier. We studied cell-type specific adaptations of human skeletal muscle to acute endurance exercise using single-nucleus (sn) multiome sequencing in human vastus lateralis samples collected before and 3.5 hours after 40 min exercise at 70% VO2max in four subjects, as well as in matched time of day samples from two supine resting circadian controls. High quality same-cell RNA-seq and ATAC-seq data were obtained from 37,154 nuclei comprising 14 cell types. Among muscle fiber types, both shared and fiber-type specific regulatory programs were identified. Single-cell circuit analysis identified distinct adaptations in fast, slow and intermediate fibers as well as LUM-expressing FAP cells, involving a total of 328 transcription factors (TFs) acting at altered accessibility sites regulating 2,025 genes. These data and circuit mapping provide single-cell insight into the processes underlying tissue and metabolic remodeling responses to exercise.

Bulk brain tissue cell-type deconvolution with bias correction for single-nuclei RNA sequencing data using DeTREM

BACKGROUND

Quantifying cell-type abundance in bulk tissue RNA-sequencing enables researchers to better understand complex systems. Newer deconvolution methodologies, such as MuSiC, use cell-type signatures derived from single-cell RNA-sequencing (scRNA-seq) data to make these calculations. Single-nuclei RNA-sequencing (snRNA-seq) reference data can be used instead of scRNA-seq data for tissues such as human brain where single-cell data are difficult to obtain, but accuracy suffers due to sequencing differences between the technologies.

RESULTS

We propose a modification to MuSiC entitled 'DeTREM' which compensates for sequencing differences between the cell-type signature and bulk RNA-seq datasets in order to better predict cell-type fractions. We show DeTREM to be more accurate than MuSiC in simulated and real human brain bulk RNA-sequencing datasets with various cell-type abundance estimates. We also compare DeTREM to SCDC and CIBERSORTx, two recent deconvolution methods that use scRNA-seq cell-type signatures. We find that they perform well in simulated data but produce less accurate results than DeTREM when used to deconvolute human brain data.

CONCLUSION

DeTREM improves the deconvolution accuracy of MuSiC and outperforms other deconvolution methods when applied to snRNA-seq data. DeTREM enables accurate cell-type deconvolution in situations where scRNA-seq data are not available. This modification improves characterization cell-type specific effects in brain tissue and identification of cell-type abundance differences under various conditions.

Projecting genetic associations through gene expression patterns highlights disease etiology and drug mechanisms

Genes act in concert with each other in specific contexts to perform their functions. Determining how these genes influence complex traits requires a mechanistic understanding of expression regulation across different conditions. It has been shown that this insight is critical for developing new therapies. Transcriptome-wide association studies have helped uncover the role of individual genes in disease-relevant mechanisms. However, modern models of the architecture of complex traits predict that gene-gene interactions play a crucial role in disease origin and progression. Here we introduce PhenoPLIER, a computational approach that maps gene-trait associations and pharmacological perturbation data into a common latent representation for a joint analysis. This representation is based on modules of genes with similar expression patterns across the same conditions. We observe that diseases are significantly associated with gene modules expressed in relevant cell types, and our approach is accurate in predicting known drug-disease pairs and inferring mechanisms of action. Furthermore, using a CRISPR screen to analyze lipid regulation, we find that functionally important players lack associations but are prioritized in trait-associated modules by PhenoPLIER. By incorporating groups of co-expressed genes, PhenoPLIER can contextualize genetic associations and reveal potential targets missed by single-gene strategies.

MousiPLIER: A Mouse Pathway-Level Information Extractor Model

High throughput gene expression profiling is a powerful approach to generate hypotheses on the underlying causes of biological function and disease. Yet this approach is limited by its ability to infer underlying biological pathways and burden of testing tens of thousands of individual genes. Machine learning models that incorporate prior biological knowledge are necessary to extract meaningful pathways and generate rational hypothesis from the vast amount of gene expression data generated to date. We adopted an unsupervised machine learning method, Pathway-level information extractor (PLIER), to train the first mouse PLIER model on 190,111 mouse brain RNA-sequencing samples, the greatest amount of training data ever used by PLIER. mousiPLER converted gene expression data into a latent variables that align to known pathway or cell maker gene sets, substantially reducing data dimensionality and improving interpretability. To determine the utility of mousiPLIER, we applied it to a mouse brain aging study of microglia and astrocyte transcriptomic profiling. We found a specific set of latent variables that are significantly associated with aging, including one latent variable (LV41) corresponding to striatal signal. We next performed k-means clustering on the training data to identify studies that respond strongly to LV41, finding that the variable is relevant to striatum and aging across the scientific literature. Finally, we built a web server (http://mousiplier.greenelab.com/) for users to easily explore the learned latent variables. Taken together this study provides proof of concept that mousiPLIER can uncover meaningful biological processes in mouse transcriptomic studies.

PINNet: a deep neural network with pathway prior knowledge for Alzheimer's disease

Introduction

Identification of Alzheimer's Disease (AD)-related transcriptomic signatures from blood is important for early diagnosis of the disease. Deep learning techniques are potent classifiers for AD diagnosis, but most have been unable to identify biomarkers because of their lack of interpretability.

Methods

To address these challenges, we propose a pathway information-based neural network (PINNet) to predict AD patients and analyze blood and brain transcriptomic signatures using an interpretable deep learning model. PINNet is a deep neural network (DNN) model with pathway prior knowledge from either the Gene Ontology or Kyoto Encyclopedia of Genes and Genomes databases. Then, a backpropagation-based model interpretation method was applied to reveal essential pathways and genes for predicting AD.

Results

The performance of PINNet was compared with a DNN model without a pathway. Performances of PINNet outperformed or were similar to those of DNN without a pathway using blood and brain gene expressions, respectively. Moreover, PINNet considers more AD-related genes as essential features than DNN without a pathway in the learning process. Pathway analysis of protein-protein interaction modules of highly contributed genes showed that AD-related genes in blood were enriched with cell migration, PI3K-Akt, MAPK signaling, and apoptosis in blood. The pathways enriched in the brain module included cell migration, PI3K-Akt, MAPK signaling, apoptosis, protein ubiquitination, and t-cell activation.

Discussion

By integrating prior knowledge about pathways, PINNet can reveal essential pathways related to AD. The source codes are available at https://github.com/DMCB-GIST/PINNet.

Deep transfer learning for clinical decision-making based on high-throughput data: comprehensive survey with benchmark results

The rapid growth of omics-based data has revolutionized biomedical research and precision medicine, allowing machine learning models to be developed for cutting-edge performance. However, despite the wealth of high-throughput data available, the performance of these models is hindered by the lack of sufficient training data, particularly in clinical research (in vivo experiments). As a result, translating this knowledge into clinical practice, such as predicting drug responses, remains a challenging task. Transfer learning is a promising tool that bridges the gap between data domains by transferring knowledge from the source to the target domain. Researchers have proposed transfer learning to predict clinical outcomes by leveraging pre-clinical data (mouse, zebrafish), highlighting its vast potential. In this work, we present a comprehensive literature review of deep transfer learning methods for health informatics and clinical decision-making, focusing on high-throughput molecular data. Previous reviews mostly covered image-based transfer learning works, while we present a more detailed analysis of transfer learning papers. Furthermore, we evaluated original studies based on different evaluation settings across cross-validations, data splits and model architectures. The result shows that those transfer learning methods have great potential; high-throughput sequencing data and state-of-the-art deep learning models lead to significant insights and conclusions. Additionally, we explored various datasets in transfer learning papers with statistics and visualization.

Machine learning in rare disease

High-throughput profiling methods (such as genomics or imaging) have accelerated basic research and made deep molecular characterization of patient samples routine. These approaches provide a rich portrait of genes, molecular pathways and cell types involved in disease phenotypes. Machine learning (ML) can be a useful tool for extracting disease-relevant patterns from high-dimensional datasets. However, depending upon the complexity of the biological question, machine learning often requires many samples to identify recurrent and biologically meaningful patterns. Rare diseases are inherently limited in clinical cases, leading to few samples to study. In this Perspective, we outline the challenges and emerging solutions for using ML for small sample sets, specifically in rare diseases. Advances in ML methods for rare diseases are likely to be informative for applications beyond rare diseases for which few samples exist with high-dimensional data. We propose that the method community prioritize the development of ML techniques for rare disease research.

Heterologous vaccination utilizing viral vector and protein platforms confers complete protection against SFTSV

Severe fever with thrombocytopenia syndrome virus was first discovered in 2009 as the causative agent of severe fever with thrombocytopenia syndrome. Despite its potential threat to public health, no prophylactic vaccine is yet available. This study developed a heterologous prime-boost strategy comprising priming with recombinant replication-deficient human adenovirus type 5 (rAd5) expressing the surface glycoprotein, Gn, and boosting with Gn protein. This vaccination regimen induced balanced Th1/Th2 immune responses and resulted in potent humoral and T cell-mediated responses in mice. It elicited high neutralizing antibody titers in both mice and non-human primates. Transcriptome analysis revealed that rAd5 and Gn proteins induced adaptive and innate immune pathways, respectively. This study provides immunological and mechanistic insight into this heterologous regimen and paves the way for future strategies against emerging infectious diseases.

Dynamic transcriptomic responses to divergent acute exercise stimuli in young adults

Acute exercise elicits dynamic transcriptional changes that, when repeated, form the fundamental basis of health, resilience, and performance adaptations. While moderate-intensity endurance training combined with conventional resistance training (traditional, TRAD) is often prescribed and recommended by public health guidance, high-intensity training combining maximal-effort intervals with intensive, limited-rest resistance training is a time-efficient alternative that may be used tactically (HITT) to confer similar benefits. Mechanisms of action of these distinct stimuli are incompletely characterized and have not been directly compared. We assessed transcriptome-wide responses in skeletal muscle and circulating extracellular vesicles (EVs) to a single exercise bout in young adults randomized to TRAD (n = 21, 12 M/9 F, 22 ± 3 yr) or HITT (n = 19, 11 M/8 F, 22 ± 2 yr). Next-generation sequencing captured small, long, and circular RNA in muscle and EVs. Analysis identified differentially expressed transcripts (|log2FC|>1, FDR ≤ 0.05) immediately (h0, EVs only), h3, and h24 postexercise within and between exercise protocols. In aaddition, all apparently responsive transcripts (FDR < 0.2) underwent singular value decomposition to summarize data structures into latent variables (LVs) to deconvolve molecular expression circuits and interregulatory relationships. LVs were compared across time and exercise protocol. TRAD, a longer but less intense stimulus, generally elicited a stronger transcriptional response than HITT, but considerable overlap and key differences existed. Findings reveal shared and unique molecular responses to the exercise stimuli and lay groundwork toward establishing relationships between protein-coding genes and lesser-understood transcripts that serve regulatory roles following exercise. Future work should advance the understanding of these circuits and whether they repeat in other populations or following other types of exercise/stress.NEW & NOTEWORTHY We examined small and long transcriptomics in skeletal muscle and serum-derived extracellular vesicles before and after a single exposure to traditional combined exercise (TRAD) and high-intensity tactical training (HITT). Across 40 young adults, we found more consistent protein-coding gene responses to TRAD, whereas HITT elicited differential expression of microRNA enriched in brain regions. Follow-up analysis revealed relationships and temporal dynamics across transcript networks, highlighting potential avenues for research into mechanisms of exercise response and adaptation.

Cross-platform normalization enables machine learning model training on microarray and RNA-seq data simultaneously

Large compendia of gene expression data have proven valuable for the discovery of novel biological relationships. Historically, most available RNA assays were run on microarray, while RNA-seq is now the platform of choice for many new experiments. The data structure and distributions between the platforms differ, making it challenging to combine them directly. Here we perform supervised and unsupervised machine learning evaluations to assess which existing normalization methods are best suited for combining microarray and RNA-seq data. We find that quantile and Training Distribution Matching normalization allow for supervised and unsupervised model training on microarray and RNA-seq data simultaneously. Nonparanormal normalization and z-scores are also appropriate for some applications, including pathway analysis with Pathway-Level Information Extractor (PLIER). We demonstrate that it is possible to perform effective cross-platform normalization using existing methods to combine microarray and RNA-seq data for machine learning applications.

A patient advocating for transparent science in rare disease research

300 million people live with at least one of 6,000 rare diseases worldwide. However, rare disease research is not always reviewed with scrutiny, making it susceptible to what the author refers to as nontransparent science. Nontransparent science can obscure animal model flaws, misguide medicine regulators and drug developers, delay or frustrate orphan drug development, or waste limited resources for rare disease research. Flawed animal models not only lack pharmacologic relevance, but also give rise to issue of clinical translatability. Sadly, these consequences and risks are grossly overlooked. Nontransparency in science can take many forms, such as premature publication of animal models without clinically significant data, not providing corrections when flaws to the model are discovered, lack of warning of critical study limitations, missing critical control data, questionable data quality, surprising results without a sound explanation, failure to rule out potential factors which may affect study conclusions, lack of sufficient detail for others to replicate the study, dubious authorship and study accountability. Science has no boarders, neither does nontransparent science. Nontransparent science can happen irrespective of the researcher's senority, institutional affiliation or country. As a patient-turned researcher suffering from Bietti crystalline dystrophy (BCD), I use BCD as an example to analyze various forms of nontransparent science in rare disease research. This article analyzes three papers published by different research groups on Cyp4v3^-/-, high-fat diet (HFD)-Cyp4v3^-/-, and Exon1-Cyp4v3^-/- mouse models of BCD. As the discussion probes various forms of nontransparent science, the flaws of these knockout mouse models are uncovered. These mouse models do not mimic BCD in humans nor do they address the lack of Cyp4v3 (murine ortholog of human CYP4V2) expression in wild type (WT) mouse retina which is markedly different from CYP4V2 expression in human retina. Further, this article discusses the impact of nontransparent science on drug development which can lead to significant delays ultimately affecting the patients. Lessons from BCD research can be helpful to all those suffering from rare diseases. As a patient, I call for transparent science in rare disease research.

Muscle transcriptomic circuits linked to periarticular physiology in end-stage osteoarthritis

The ability of individuals with end-stage osteoarthritis (OA) to functionally recover from total joint arthroplasty is highly inconsistent. The molecular mechanisms driving this heterogeneity have yet to be elucidated. Furthermore, OA disproportionately impacts females, suggesting a need for identifying female-specific therapeutic targets. We profiled the skeletal muscle transcriptome in females with end-stage OA (n = 20) undergoing total knee or hip arthroplasty using RNA-Seq. Single-gene differential expression (DE) analyses tested for DE genes between skeletal muscle overlaying the surgical (SX) joint and muscle from the contralateral (CTRL) leg. Network analyses were performed using Pathway-Level Information ExtractoR (PLIER) to summarize genes into latent variables (LVs), i.e., gene circuits, and link them to biological pathways. LV differences in SX versus CTRL muscle and across sources of muscle tissue (vastus medialis, vastus lateralis, or tensor fascia latae) were determined with ANOVA. Linear models tested for associations between LVs and muscle phenotype on the SX side (inflammation, function, and integrity). DE analysis revealed 360 DE genes (|Log2 fold-difference| ≥ 1, FDR ≤ 0.05) between the SX and CTRL limbs, many associated with inflammation and lipid metabolism. PLIER analyses revealed circuits associated with protein degradation and fibro-adipogenic cell gene expression. Muscle inflammation and function were linked to an LV associated with endothelial cell gene expression highlighting a potential regulatory role of endothelial cells within skeletal muscle. These findings may provide insight into potential therapeutic targets to improve OA rehabilitation before and/or following total joint replacement.

Pre-infection antiviral innate immunity contributes to sex differences in SARS-CoV-2 infection

Male sex is a major risk factor for SARS-CoV-2 infection severity. To understand the basis for this sex difference, we studied SARS-CoV-2 infection in a young adult cohort of United States Marine recruits. Among 2,641 male and 244 female unvaccinated and seronegative recruits studied longitudinally, SARS-CoV-2 infections occurred in 1,033 males and 137 females. We identified sex differences in symptoms, viral load, blood transcriptome, RNA splicing, and proteomic signatures. Females had higher pre-infection expression of antiviral interferon-stimulated gene (ISG) programs. Causal mediation analysis implicated ISG differences in number of symptoms, levels of ISGs, and differential splicing of CD45 lymphocyte phosphatase during infection. Our results indicate that the antiviral innate immunity set point causally contributes to sex differences in response to SARS-CoV-2 infection. A record of this paper's transparent peer review process is included in the supplemental information.

Proteogenomics reveals sex-biased aging genes and coordinated splicing in cardiac aging

The risks of heart diseases are significantly modulated by age and sex, but how these factors influence baseline cardiac gene expression remains incompletely understood. Here, we used RNA sequencing and mass spectrometry to compare gene expression in female and male young adult (4 mo) and early aging (20 mo) mouse hearts, identifying thousands of age- and sex-dependent gene expression signatures. Sexually dimorphic cardiac genes are broadly distributed, functioning in mitochondrial metabolism, translation, and other processes. In parallel, we found over 800 genes with differential aging response between male and female, including genes in cAMP and PKA signaling. Analysis of the sex-adjusted aging cardiac transcriptome revealed a widespread remodeling of exon usage patterns that is largely independent from differential gene expression, concomitant with upstream changes in RNA-binding protein and splice factor transcripts. To evaluate the impact of the splicing events on cardiac proteoform composition, we applied an RNA-guided proteomics computational pipeline to analyze the mass spectrometry data and detected hundreds of putative splice variant proteins that have the potential to rewire the cardiac proteome. Taken together, the results here suggest that cardiac aging is associated with 1) widespread sex-biased aging genes and 2) a rewiring of RNA splicing programs, including sex- and age-dependent changes in exon usages and splice patterns that have the potential to influence cardiac protein structure and function. These changes contribute to the emerging evidence for considerable sexual dimorphism in the cardiac aging process that should be considered in the search for disease mechanisms.NEW & NOTEWORTHY Han et al. used proteogenomics to compare male and female mouse hearts at 4 and 20 mo. Sex-biased cardiac genes function in mitochondrial metabolism, translation, autophagy, and other processes. Hundreds of cardiac genes show sex-by-age interactions, that is, sex-biased aging genes. Cardiac aging is accompanied with a remodeling of exon usage in functionally coordinated genes, concomitant with differential expression of RNA-binding proteins and splice factors. These features represent an underinvestigated aspect of cardiac aging that may be relevant to the search for disease mechanisms.

A benchmark study of deep learning-based multi-omics data fusion methods for cancer

BACKGROUND

A fused method using a combination of multi-omics data enables a comprehensive study of complex biological processes and highlights the interrelationship of relevant biomolecules and their functions. Driven by high-throughput sequencing technologies, several promising deep learning methods have been proposed for fusing multi-omics data generated from a large number of samples.

RESULTS

In this study, 16 representative deep learning methods are comprehensively evaluated on simulated, single-cell, and cancer multi-omics datasets. For each of the datasets, two tasks are designed: classification and clustering. The classification performance is evaluated by using three benchmarking metrics including accuracy, F1 macro, and F1 weighted. Meanwhile, the clustering performance is evaluated by using four benchmarking metrics including the Jaccard index (JI), C-index, silhouette score, and Davies Bouldin score. For the cancer multi-omics datasets, the methods' strength in capturing the association of multi-omics dimensionality reduction results with survival and clinical annotations is further evaluated. The benchmarking results indicate that moGAT achieves the best classification performance. Meanwhile, efmmdVAE, efVAE, and lfmmdVAE show the most promising performance across all complementary contexts in clustering tasks.

CONCLUSIONS

Our benchmarking results not only provide a reference for biomedical researchers to choose appropriate deep learning-based multi-omics data fusion methods, but also suggest the future directions for the development of more effective multi-omics data fusion methods. The deep learning frameworks are available at https://github.com/zhenglinyi/DL-mo .

Pyramidal neuron subtype diversity governs microglia states in the neocortex

Microglia are specialized macrophages in the brain parenchyma that exist in multiple transcriptional states and reside within a wide range of neuronal environments1-4. However, how and where these states are generated remains poorly understood. Here, using the mouse somatosensory cortex, we demonstrate that microglia density and molecular state acquisition are determined by the local composition of pyramidal neuron classes. Using single-cell and spatial transcriptomic profiling, we unveil the molecular signatures and spatial distributions of diverse microglia populations and show that certain states are enriched in specific cortical layers, whereas others are broadly distributed throughout the cortex. Notably, conversion of deep-layer pyramidal neurons to an alternate class identity reconfigures the distribution of local, layer-enriched homeostatic microglia to match the new neuronal niche. Leveraging the transcriptional diversity of pyramidal neurons in the neocortex, we construct a ligand-receptor atlas describing interactions between individual pyramidal neuron subtypes and microglia states, revealing rules of neuron-microglia communication. Our findings uncover a fundamental role for neuronal diversity in instructing the acquisition of microglia states as a potential mechanism for fine-tuning neuroimmune interactions within the cortical local circuitry.

GenomicSuperSignature facilitates interpretation of RNA-seq experiments through robust, efficient comparison to public databases

Millions of transcriptomic profiles have been deposited in public archives, yet remain underused for the interpretation of new experiments. We present a method for interpreting new transcriptomic datasets through instant comparison to public datasets without high-performance computing requirements. We apply Principal Component Analysis on 536 studies comprising 44,890 human RNA sequencing profiles and aggregate sufficiently similar loading vectors to form Replicable Axes of Variation (RAV). RAVs are annotated with metadata of originating studies and by gene set enrichment analysis. Functionality to associate new datasets with RAVs, extract interpretable annotations, and provide intuitive visualization are implemented as the GenomicSuperSignature R/Bioconductor package. We demonstrate the efficient and coherent database search, robustness to batch effects and heterogeneous training data, and transfer learning capacity of our method using TCGA and rare diseases datasets. GenomicSuperSignature aids in analyzing new gene expression data in the context of existing databases using minimal computing resources.

Learning biologically-interpretable latent representations for gene expression data: Pathway Activity Score Learning Algorithm

Molecular gene-expression datasets consist of samples with tens of thousands of measured quantities (i.e., high dimensional data). However, lower-dimensional representations that retain the useful biological information do exist. We present a novel algorithm for such dimensionality reduction called Pathway Activity Score Learning (PASL). The major novelty of PASL is that the constructed features directly correspond to known molecular pathways (genesets in general) and can be interpreted as pathway activity scores. Hence, unlike PCA and similar methods, PASL's latent space has a fairly straightforward biological interpretation. PASL is shown to outperform in predictive performance the state-of-the-art method (PLIER) on two collections of breast cancer and leukemia gene expression datasets. PASL is also trained on a large corpus of 50000 gene expression samples to construct a universal dictionary of features across different tissues and pathologies. The dictionary validated on 35643 held-out samples for reconstruction error. It is then applied on 165 held-out datasets spanning a diverse range of diseases. The AutoML tool JADBio is employed to show that the predictive information in the PASL-created feature space is retained after the transformation. The code is available at https://github.com/mensxmachina/PASL.

A single intranasal dose of human parainfluenza virus type 3-vectored vaccine induces effective antibody and memory T cell response in the lungs and protects hamsters against SARS-CoV-2

Respiratory tract vaccination has an advantage of needle-free delivery and induction of mucosal immune response in the portal of SARS-CoV-2 entry. We utilized human parainfluenza virus type 3 vector to generate constructs expressing the full spike (S) protein of SARS-CoV-2, its S1 subunit, or the receptor-binding domain, and tested them in hamsters as single-dose intranasal vaccines. The construct bearing full-length S induced high titers of neutralizing antibodies specific to S protein domains critical to the protein functions. Robust memory T cell responses in the lungs were also induced, which represent an additional barrier to infection and should be less sensitive than the antibody responses to mutations present in SARS-CoV-2 variants. Following SARS-CoV-2 challenge, animals were protected from the disease and detectable viral replication. Vaccination prevented induction of gene pathways associated with inflammation. These results indicate advantages of respiratory vaccination against COVID-19 and inform the design of mucosal SARS-CoV-2 vaccines.

Transfer learning of clinical outcomes from preclinical molecular data, principles and perspectives

Accurate transfer learning of clinical outcomes from one cellular context to another, between cell types, developmental stages, omics modalities or species, is considered tremendously useful. When transferring a prediction task from a source domain to a target domain, what counts is the high quality of the predictions in the target domain, requiring states or processes common to both the source and the target that can be learned by the predictor reflected by shared denominators. These may form a compendium of knowledge that is learned in the source to enable predictions in the target, usually with few, if any, labeled target training samples to learn from. Transductive transfer learning refers to the learning of the predictor in the source domain, transferring its outcome label calculations to the target domain, considering the same task. Inductive transfer learning considers cases where the target predictor is performing a different yet related task as compared with the source predictor. Often, there is also a need to first map the variables in the input/feature spaces and/or the variables in the output/outcome spaces. We here discuss and juxtapose various recently published transfer learning approaches, specifically designed (or at least adaptable) to predict clinical (human in vivo) outcomes based on preclinical (mostly animal-based) molecular data, towards finding the right tool for a given task, and paving the way for a comprehensive and systematic comparison of the suitability and accuracy of transfer learning of clinical outcomes.

Latent Variables Capture Pathway-Level Points of Departure in High-Throughput Toxicogenomic Data

Estimation of points of departure (PoDs) from high-throughput transcriptomic data (HTTr) represents a key step in the development of next-generation risk assessment (NGRA). Current approaches mainly rely on single key gene targets, which are constrained by the information currently available in the knowledge base and make interpretation challenging as scientists need to interpret PoDs for thousands of genes or hundreds of pathways. In this work, we aimed to address these issues by developing a computational workflow to investigate the pathway concentration–response relationships in a way that is not fully constrained by known biology and also facilitates interpretation. We employed the Pathway-Level Information ExtractoR (PLIER) to identify latent variables (LVs) describing biological activity and then investigated in vitro LVs’ concentration–response relationships using the ToxCast pipeline. We applied this methodology to a published transcriptomic concentration–response data set for 44 chemicals in MCF-7 cells and showed that our workflow can capture known biological activity and discriminate between estrogenic and antiestrogenic compounds as well as activity not aligning with the existing knowledge base, which may be relevant in a risk assessment scenario. Moreover, we were able to identify the known estrogen activity in compounds that are not well-established ER agonists/antagonists supporting the use of the workflow in read-across. Next, we transferred its application to chemical compounds tested in HepG2, HepaRG, and MCF-7 cells and showed that PoD estimates are in strong agreement with those estimated using a recently developed Bayesian approach (cor = 0.89) and in weak agreement with those estimated using a well-established approach such as BMDExpress2 (cor = 0.57). These results demonstrate the effectiveness of using PLIER in a concentration–response scenario to investigate pathway activity in a way that is not fully constrained by the knowledge base and to ease the biological interpretation and support the development of an NGRA framework with the ability to improve current risk assessment strategies for chemicals using new approach methodologies.

Single nucleus transcriptome and chromatin accessibility of postmortem human pituitaries reveal diverse stem cell regulatory mechanisms

Despite their importance in tissue homeostasis and renewal, human pituitary stem cells (PSCs) are incompletely characterized. We describe a human single nucleus RNA-seq and ATAC-seq resource from pediatric, adult, and aged postmortem pituitaries (snpituitaryatlas.princeton.edu) and characterize cell-type-specific gene expression and chromatin accessibility programs for all major pituitary cell lineages. We identify uncommitted PSCs, committing progenitor cells, and sex differences. Pseudotime trajectory analysis indicates that early-life PSCs are distinct from the other age groups. Linear modeling of same-cell multiome data identifies regulatory domain accessibility sites and transcription factors that are significantly associated with gene expression in PSCs compared with other cell types and within PSCs. We identify distinct deterministic mechanisms that contribute to heterogeneous marker expression within PSCs. These findings characterize human stem cell lineages and reveal diverse mechanisms regulating key PSC genes and cell type identity.

This study profiles the gene expression and chromatin accessibility landscapes in postmortem male and female pituitaries of different ages using single nucleus multiomics technologies. Zhang et al. characterize the pituitary stem cell population and develop computational methods, which allow us to elucidate regulatory mechanisms underlying pituitary stem cell identity.

Skeletal muscle transcriptome response to a bout of endurance exercise in physically active and sedentary older adults

Age-related declines in cardiorespiratory fitness and physical function are mitigated by regular endurance exercise in older adults. This may be due, in part, to changes in the transcriptional program of skeletal muscle following repeated bouts of exercise. However, the impact of chronic exercise training on the transcriptional response to an acute bout of endurance exercise has not been clearly determined. Here, we characterized baseline differences in muscle transcriptome and exercise-induced response in older adults who were active/endurance trained or sedentary. RNA-sequencing was performed on vastus lateralis biopsy specimens obtained before, immediately after, and 3 h following a bout of endurance exercise (40 min of cycling at 60%-70% of heart rate reserve). Using a recently developed bioinformatics approach, we found that transcript signatures related to type I myofibers, mitochondria, and endothelial cells were higher in active/endurance-trained adults and were associated with key phenotypic features including V̇o_2peak, ATP_max, and muscle fiber proportion. Immune cell signatures were elevated in the sedentary group and linked to visceral and intermuscular adipose tissue mass. Following acute exercise, we observed distinct temporal transcriptional signatures that were largely similar among groups. Enrichment analysis revealed catabolic processes were uniquely enriched in the sedentary group at the 3-h postexercise timepoint. In summary, this study revealed key transcriptional signatures that distinguished active and sedentary adults, which were associated with difference in oxidative capacity and depot-specific adiposity. The acute response signatures were consistent with beneficial effects of endurance exercise to improve muscle health in older adults irrespective of exercise history and adiposity.NEW & NOTEWORTHY Muscle transcript signatures associated with oxidative capacity and immune cells underlie important phenotypic and clinical characteristics of older adults who are endurance trained or sedentary. Despite divergent phenotypes, the temporal transcriptional signatures in response to an acute bout of endurance exercise were largely similar among groups. These data provide new insight into the transcriptional programs of aging muscle and the beneficial effects of endurance exercise to promote healthy aging in older adults.

Facioscapulohumeral dystrophy transcriptome signatures correlate with different stages of disease and are marked by different MRI biomarkers

With several therapeutic strategies for facioscapulohumeral muscular dystrophy (FSHD) entering clinical testing, outcome measures are becoming increasingly important. Considering the spatiotemporal nature of FSHD disease activity, clinical trials would benefit from non-invasive imaging-based biomarkers that can predict FSHD-associated transcriptome changes. This study investigated two FSHD-associated transcriptome signatures (DUX4 and PAX7 signatures) in FSHD skeletal muscle biopsies, and tested their correlation with a variety of disease-associated factors, including Ricci clinical severity score, disease duration, D4Z4 repeat size, muscle pathology scorings and functional outcome measures. It establishes that DUX4 and PAX7 signatures both show a sporadic expression pattern in FSHD-affected biopsies, possibly marking different stages of disease. This study analyzed two imaging-based biomarkers—Turbo Inversion Recovery Magnitude (TIRM) hyperintensity and fat fraction—and provides insights into their predictive power as non-invasive biomarkers for FSHD signature detection in clinical trials. Further insights in the heterogeneity of—and correlation between—imaging biomarkers and molecular biomarkers, as provided in this study, will provide important guidance to clinical trial design in FSHD. Finally, this study investigated the role of infiltrating non-muscle cell types in FSHD signature expression and detected potential distinct roles for two fibro-adipogenic progenitor subtypes in FSHD.

Cascading epigenomic analysis for identifying disease genes from the regulatory landscape of GWAS variants

The majority of genetic variants detected in genome wide association studies (GWAS) exert their effects on phenotypes through gene regulation. Motivated by this observation, we propose a multi-omic integration method that models the cascading effects of genetic variants from epigenome to transcriptome and eventually to the phenome in identifying target genes influenced by risk alleles. This cascading epigenomic analysis for GWAS, which we refer to as CEWAS, comprises two types of models: one for linking cis genetic effects to epigenomic variation and another for linking cis epigenomic variation to gene expression. Applying these models in cascade to GWAS summary statistics generates gene level statistics that reflect genetically-driven epigenomic effects. We show on sixteen brain-related GWAS that CEWAS provides higher gene detection rate than related methods, and finds disease relevant genes and gene sets that point toward less explored biological processes. CEWAS thus presents a novel means for exploring the regulatory landscape of GWAS variants in uncovering disease mechanisms.

The majority of genetic variants detected in genome wide association studies (GWAS) exert their effects on phenotypes through gene regulation. Motivated by this observation, we propose a multi-omic integration method that models the cascading effects of genetic variants from epigenome to transcriptome and eventually to the phenome in identifying target genes influenced by risk alleles. This cascading epigenomic analysis for GWAS, which we refer to as CEWAS, combines the effect of genetic variants on DNA methylation as well as gene expression. We show on sixteen brain-related GWAS that CEWAS provides higher gene detection rate than related methods, and finds disease relevant genes and gene sets that point toward less explored biological processes.

Forecasting cellular states: from descriptive to predictive biology via single-cell multiomics

As the single cell field races to characterize each cell type, state, and behavior, the complexity of the computational analysis approaches the complexity of the biological systems. Single cell and imaging technologies now enable unprecedented measurements of state transitions in biological systems, providing high-throughput data that capture tens-of-thousands of measurements on hundreds-of-thousands of samples. Thus, the definition of cell type and state is evolving to encompass the broad range of biological questions now attainable. To answer these questions requires the development of computational tools for integrated multi-omics analysis. Merged with mathematical models, these algorithms will be able to forecast future states of biological systems, going from statistical inferences of phenotypes to time course predictions of the biological systems with dynamic maps analogous to weather systems. Thus, systems biology for forecasting biological system dynamics from multi-omic data represents the future of cell biology empowering a new generation of technology-driven predictive medicine.

Single nucleus multi-omics regulatory landscape of the murine pituitary

To provide a multi-omics resource and investigate transcriptional regulatory mechanisms, we profile the transcriptome, chromatin accessibility, and methylation status of over 70,000 single nuclei (sn) from adult mouse pituitaries. Paired snRNAseq and snATACseq datasets from individual animals highlight a continuum between developmental epigenetically-encoded cell types and transcriptionally-determined transient cell states. Co-accessibility analysis-based identification of a putative Fshb cis-regulatory domain that overlaps the fertility-linked rs11031006 human polymorphism, followed by experimental validation illustrate the use of this resource for hypothesis generation. We also identify transcriptional and chromatin accessibility programs distinguishing each major cell type. Regulons, which are co-regulated gene sets sharing binding sites for a common transcription factor driver, recapitulate cell type clustering. We identify both cell type-specific and sex-specific regulons that are highly correlated with promoter accessibility, but not with methylation state, supporting the centrality of chromatin accessibility in shaping cell-defining transcriptional programs. The sn multi-omics atlas is accessible at snpituitaryatlas.princeton.edu.

Muscle transcriptional networks linked to resistance exercise training hypertrophic response heterogeneity

The skeletal muscle hypertrophic response to resistance exercise training (RT) is highly variable across individuals. The molecular underpinnings of this heterogeneity are unclear. This study investigated transcriptional networks linked to RT-induced muscle hypertrophy, classified as 1) predictive of hypertrophy, 2) responsive to RT independent of muscle hypertrophy, or 3) plastic with hypertrophy. Older adults (n = 31, 18 F/13 M, 70 ± 4 yr) underwent 14-wk RT (3 days/wk, alternating high-low-high intensity). Muscle hypertrophy was assessed by pre- to post-RT change in mid-thigh muscle cross-sectional area (CSA) [computed tomography (CT), primary outcome] and thigh lean mass [dual-energy X-ray absorptiometry (DXA), secondary outcome]. Transcriptome-wide poly-A RNA-seq was performed on vastus lateralis tissue collected pre- (n = 31) and post-RT (n = 22). Prediction networks (using only baseline RNA-seq) were identified by weighted gene correlation network analysis (WGCNA). To identify Plasticity networks, WGCNA change indices for paired samples were calculated and correlated to changes in muscle size outcomes. Pathway-level information extractor (PLIER) was applied to identify Response networks and link genes to biological annotation. Prediction networks (n = 6) confirmed transcripts previously connected to resistance/aerobic training adaptations in the MetaMEx database while revealing novel member genes that should fuel future research to understand the influence of baseline muscle gene expression on hypertrophy. Response networks (n = 6) indicated RT-induced increase in aerobic metabolism and reduced expression of genes associated with spliceosome biology and type-I myofibers. A single exploratory Plasticity network was identified. Findings support that interindividual differences in baseline gene expression may contribute more than RT-induced changes in gene networks to muscle hypertrophic response heterogeneity. Code/Data: https://github.com/kallavin/MASTERS_manuscript/tree/master.

Parameterized algorithms for identifying gene co-expression modules via weighted clique decomposition

We present a new combinatorial model for identifying regulatory modules in gene co-expression data using a decomposition into weighted cliques. To capture complex interaction effects, we generalize the previously-studied weighted edge clique partition problem. As a first step, we restrict ourselves to the noise-free setting, and show that the problem is fixed parameter tractable when parameterized by the number of modules (cliques). We present two new algorithms for finding these decompositions, using linear programming and integer partitioning to determine the clique weights. Further, we implement these algorithms in Python and test them on a biologically-inspired synthetic corpus generated using real-world data from transcription factors and a latent variable analysis of co-expression in varying cell types.

Integrated multi-omics approaches to improve classification of chronic kidney disease

Chronic kidney diseases (CKDs) are currently classified according to their clinical features, associated comorbidities and pattern of injury on biopsy. Even within a given classification, considerable variation exists in disease presentation, progression and response to therapy, highlighting heterogeneity in the underlying biological mechanisms. As a result, patients and clinicians experience uncertainty when considering optimal treatment approaches and risk projection. Technological advances now enable large-scale datasets, including DNA and RNA sequence data, proteomics and metabolomics data, to be captured from individuals and groups of patients along the genotype-phenotype continuum of CKD. The ability to combine these high-dimensional datasets, in which the number of variables exceeds the number of clinical outcome observations, using computational approaches such as machine learning, provides an opportunity to re-classify patients into molecularly defined subgroups that better reflect underlying disease mechanisms. Patients with CKD are uniquely poised to benefit from these integrative, multi-omics approaches since the kidney biopsy, blood and urine samples used to generate these different types of molecular data are frequently obtained during routine clinical care. The ultimate goal of developing an integrated molecular classification is to improve diagnostic classification, risk stratification and assignment of molecular, disease-specific therapies to improve the care of patients with CKD.

Co-expression analysis reveals interpretable gene modules controlled by trans-acting genetic variants

Understanding the causal processes that contribute to disease onset and progression is essential for developing novel therapies. Although trans-acting expression quantitative trait loci (trans-eQTLs) can directly reveal cellular processes modulated by disease variants, detecting trans-eQTLs remains challenging due to their small effect sizes. Here, we analysed gene expression and genotype data from six blood cell types from 226 to 710 individuals. We used co-expression modules inferred from gene expression data with five methods as traits in trans-eQTL analysis to limit multiple testing and improve interpretability. In addition to replicating three established associations, we discovered a novel trans-eQTL near SLC39A8 regulating a module of metallothionein genes in LPS-stimulated monocytes. Interestingly, this effect was mediated by a transient cis-eQTL present only in early LPS response and lost before the trans effect appeared. Our analyses highlight how co-expression combined with functional enrichment analysis improves the identification and prioritisation of trans-eQTLs when applied to emerging cell-type-specific datasets.

Deconvolving the contributions of cell-type heterogeneity on cortical gene expression

Complexity of cell-type composition has created much skepticism surrounding the interpretation of bulk tissue transcriptomic studies. Recent studies have shown that deconvolution algorithms can be applied to computationally estimate cell-type proportions from gene expression data of bulk blood samples, but their performance when applied to brain tissue is unclear. Here, we have generated an immunohistochemistry (IHC) dataset for five major cell-types from brain tissue of 70 individuals, who also have bulk cortical gene expression data. With the IHC data as the benchmark, this resource enables quantitative assessment of deconvolution algorithms for brain tissue. We apply existing deconvolution algorithms to brain tissue by using marker sets derived from human brain single cell and cell-sorted RNA-seq data. We show that these algorithms can indeed produce informative estimates of constituent cell-type proportions. In fact, neuronal subpopulations can also be estimated from bulk brain tissue samples. Further, we show that including the cell-type proportion estimates as confounding factors is important for reducing false associations between Alzheimer's disease phenotypes and gene expression. Lastly, we demonstrate that using more accurate marker sets can substantially improve statistical power in detecting cell-type specific expression quantitative trait loci (eQTLs).

Rehabilitative Impact of Exercise Training on Human Skeletal Muscle Transcriptional Programs in Parkinson's Disease

Parkinson's disease (PD) is the most common motor neurodegenerative disease, and neuromuscular function deficits associated with PD contribute to disability. Targeting these symptoms, our laboratory has previously evaluated 16-week high-intensity resistance exercise as rehabilitative training (RT) in individuals with PD. We reported significant improvements in muscle mass, neuromuscular function (strength, power, and motor unit activation), indices of neuromuscular junction integrity, total and motor scores on the unified Parkinson's disease rating scale (UPDRS), and total and sub-scores on the 39-item PD Quality of Life Questionnaire (PDQ-39), supporting the use of RT to reverse symptoms. Our objective was to identify transcriptional networks that may contribute to RT-induced neuromuscular remodeling in PD. We generated transcriptome-wide skeletal muscle RNA-sequencing in 5 participants with PD [4M/1F, 67 ± 2 years, Hoehn and Yahr stages 2 (n = 3) and 3 (n = 2)] before and after 16-week high intensity RT to identify transcriptional networks that may in part underpin RT-induced neuromuscular remodeling in PD. Following RT, 304 genes were significantly upregulated, notably related to remodeling and nervous system/muscle development. Additionally, 402 genes, primarily negative regulators of muscle adaptation, were downregulated. We applied the recently developed Pathway-Level Information ExtractoR (PLIER) method to reveal coordinated gene programs (as latent variables, LVs) that differed in skeletal muscle among young (YA) and old (OA) healthy adults and PD (n = 12 per cohort) at baseline and in PD pre- vs. post-RT. Notably, one LV associated with angiogenesis, axon guidance, and muscle remodeling was significantly lower in PD than YA at baseline and was significantly increased by exercise. A different LV annotated to denervation, autophagy, and apoptosis was increased in both PD and OA relative to YA and was also reduced by 16-week RT in PD. Thus, this analysis identified two novel skeletal muscle transcriptional programs that are dysregulated by PD and aging, respectively. Notably, RT has a normalizing effect on both programs in individuals with PD. These results identify potential molecular transducers of the RT-induced improvements in neuromuscular remodeling and motor function that may aid in optimizing exercise rehabilitation strategies for individuals with PD.

Integrative Analysis Identifies Candidate Tumor Microenvironment and Intracellular Signaling Pathways that Define Tumor Heterogeneity in NF1

Neurofibromatosis type 1 (NF1) is a monogenic syndrome that gives rise to numerous symptoms including cognitive impairment, skeletal abnormalities, and growth of benign nerve sheath tumors. Nearly all NF1 patients develop cutaneous neurofibromas (cNFs), which occur on the skin surface, whereas 40-60% of patients develop plexiform neurofibromas (pNFs), which are deeply embedded in the peripheral nerves. Patients with pNFs have a ~10% lifetime chance of these tumors becoming malignant peripheral nerve sheath tumors (MPNSTs). These tumors have a severe prognosis and few treatment options other than surgery. Given the lack of therapeutic options available to patients with these tumors, identification of druggable pathways or other key molecular features could aid ongoing therapeutic discovery studies. In this work, we used statistical and machine learning methods to analyze 77 NF1 tumors with genomic data to characterize key signaling pathways that distinguish these tumors and identify candidates for drug development. We identified subsets of latent gene expression variables that may be important in the identification and etiology of cNFs, pNFs, other neurofibromas, and MPNSTs. Furthermore, we characterized the association between these latent variables and genetic variants, immune deconvolution predictions, and protein activity predictions.

Incorporating biological structure into machine learning models in biomedicine

In biomedical applications of machine learning, relevant information often has a rich structure that is not easily encoded as real-valued predictors. Examples of such data include DNA or RNA sequences, gene sets or pathways, gene interaction or coexpression networks, ontologies, and phylogenetic trees. We highlight recent examples of machine learning models that use structure to constrain model architecture or incorporate structured data into model training. For machine learning in biomedicine, where sample size is limited and model interpretability is crucial, incorporating prior knowledge in the form of structured data can be particularly useful. The area of research would benefit from performant open source implementations and independent benchmarking efforts.

Single-cell transcriptional profiles in human skeletal muscle

Skeletal muscle is a heterogeneous tissue comprised of muscle fiber and mononuclear cell types that, in addition to movement, influences immunity, metabolism and cognition. We investigated the gene expression patterns of skeletal muscle cells using RNA-seq of subtype-pooled single human muscle fibers and single cell RNA-seq of mononuclear cells from human vastus lateralis, mouse quadriceps, and mouse diaphragm. We identified 11 human skeletal muscle mononuclear cell types, including two fibro-adipogenic progenitor (FAP) cell subtypes. The human FBN1+ FAP cell subtype is novel and a corresponding FBN1+ FAP cell type was also found in single cell RNA-seq analysis in mouse. Transcriptome exercise studies using bulk tissue analysis do not resolve changes in individual cell-type proportion or gene expression. The cell-type gene signatures provide the means to use computational methods to identify cell-type level changes in bulk studies. As an example, we analyzed public transcriptome data from an exercise training study and revealed significant changes in specific mononuclear cell-type proportions related to age, sex, acute exercise and training. Our single-cell expression map of skeletal muscle cell types will further the understanding of the diverse effects of exercise and the pathophysiology of muscle disease.

Pathway Activity Score Learning for Dimensionality Reduction of Gene Expression Data

Molecular gene-expression datasets consist of samples with tens of thousands of measured quantities (e.g., high dimensional data). However, there exist lower-dimensional representations that retain the useful information. We present a novel algorithm for such dimensionality reduction called Pathway Activity Score Learning (PASL). The major novelty of PASL is that the constructed features directly correspond to known molecular pathways and can be interpreted as pathway activity scores. Hence, unlike PCA and similar methods, PASL’s latent space has a relatively straight-forward biological interpretation. As a use-case, PASL is applied on two collections of breast cancer and leukemia gene expression datasets. We show that PASL does retain the predictive information for disease classification on new, unseen datasets, as well as outperforming PLIER, a recently proposed competitive method. We also show that differential activation pathway analysis provides complementary information to standard gene set enrichment analysis. The code is available at https://github.com/mensxmachina/PASL.