iGEMS: an integrated model for identification of alternative exon usage events

Transcriptomics for Clinical and Experimental Biology Research: Hang on a Seq

Sequencing the human genome empowers translational medicine, facilitating transcriptome-wide molecular diagnosis, pathway biology, and drug repositioning. Initially, microarrays are used to study the bulk transcriptome; but now short-read RNA sequencing (RNA-seq) predominates. Positioned as a superior technology, that makes the discovery of novel transcripts routine, most RNA-seq analyses are in fact modeled on the known transcriptome. Limitations of the RNA-seq methodology have emerged, while the design of, and the analysis strategies applied to, arrays have matured. An equitable comparison between these technologies is provided, highlighting advantages that modern arrays hold over RNA-seq. Array protocols more accurately quantify constitutively expressed protein coding genes across tissue replicates, and are more reliable for studying lower expressed genes. Arrays reveal long noncoding RNAs (lncRNA) are neither sparsely nor lower expressed than protein coding genes. Heterogeneous coverage of constitutively expressed genes observed with RNA-seq, undermines the validity and reproducibility of pathway analyses. The factors driving these observations, many of which are relevant to long-read or single-cell sequencing are discussed. As proposed herein, a reappreciation of bulk transcriptomic methods is required, including wider use of the modern high-density array data-to urgently revise existing anatomical RNA reference atlases and assist with more accurate study of lncRNAs.

Urgent need for consistent standards in functional enrichment analysis

Gene set enrichment tests (a.k.a. functional enrichment analysis) are among the most frequently used methods in computational biology. Despite this popularity, there are concerns that these methods are being applied incorrectly and the results of some peer-reviewed publications are unreliable. These problems include the use of inappropriate background gene lists, lack of false discovery rate correction and lack of methodological detail. To ascertain the frequency of these issues in the literature, we performed a screen of 186 open-access research articles describing functional enrichment results. We find that 95% of analyses using over-representation tests did not implement an appropriate background gene list or did not describe this in the methods. Failure to perform p-value correction for multiple tests was identified in 43% of analyses. Many studies lacked detail in the methods section about the tools and gene sets used. An extension of this survey showed that these problems are not associated with journal or article level bibliometrics. Using seven independent RNA-seq datasets, we show misuse of enrichment tools alters results substantially. In conclusion, most published functional enrichment studies suffered from one or more major flaws, highlighting the need for stronger standards for enrichment analysis.

Functional enrichment analysis is a commonly used technique to identify trends in large scale biological datasets. In biomedicine, functional enrichment analysis of gene expression data is frequently applied to identify disease and drug mechanisms. While enrichment tests were once primarily conducted with complicated computer scripts, web-based tools are becoming more widely used. Users can paste a list of genes into a website and receive enrichment results in a matter of seconds. Despite the popularity of these tools, there are concerns that statistical problems and incomplete reporting are compromising research quality. In this article, we conducted a systematic examination of published enrichment analyses and assessed whether (i) any statistical flaws were present and (ii) sufficient methodological detail is provided such that the study could be replicated. We found that lack of methodological detail and errors in statistical analysis were widespread, which undermines the reliability and reproducibility of these research articles. A set of best practices is urgently needed to raise the quality of published work.

A human-based multi-gene signature enables quantitative drug repurposing for metabolic disease

Insulin resistance (IR) contributes to the pathophysiology of diabetes, dementia, viral infection, and cardiovascular disease. Drug repurposing (DR) may identify treatments for IR; however, barriers include uncertainty whether in vitro transcriptomic assays yield quantitative pharmacological data, or how to optimise assay design to best reflect in vivo human disease. We developed a clinical-based human tissue IR signature by combining lifestyle-mediated treatment responses (>500 human adipose and muscle biopsies) with biomarkers of disease status (fasting IR from >1200 biopsies). The assay identified a chemically diverse set of >130 positively acting compounds, highly enriched in true positives, that targeted 73 proteins regulating IR pathways. Our multi-gene RNA assay score reflected the quantitative pharmacological properties of a set of epidermal growth factor receptor-related tyrosine kinase inhibitors, providing insight into drug target specificity; an observation supported by deep learning-based genome-wide predicted pharmacology. Several drugs identified are suitable for evaluation in patients, particularly those with either acute or severe chronic IR.

Human and mouse muscle transcriptomic analyses identify insulin receptor mRNA downregulation in hyperinsulinemia-associated insulin resistance

Hyperinsulinemia is commonly viewed as a compensatory response to insulin resistance, yet studies have demonstrated that chronically elevated insulin may also drive insulin resistance. The molecular mechanisms underpinning this potentially cyclic process remain poorly defined, especially on a transcriptome-wide level. Transcriptomic meta-analysis in >450 human samples demonstrated that fasting insulin reliably and negatively correlated with INSR mRNA in skeletal muscle. To establish causality and study the direct effects of prolonged exposure to excess insulin in muscle cells, we incubated C2C12 myotubes with elevated insulin for 16 h, followed by 6 h of serum starvation, and established that acute AKT and ERK signaling were attenuated in this model of in vitro hyperinsulinemia. Global RNA-sequencing of cells both before and after nutrient withdrawal highlighted genes in the insulin receptor (INSR) signaling, FOXO signaling, and glucose metabolism pathways indicative of 'hyperinsulinemia' and 'starvation' programs. Consistently, we observed that hyperinsulinemia led to a substantial reduction in Insr gene expression, and subsequently a reduced surface INSR and total INSR protein, both in vitro and in vivo. Bioinformatic modeling combined with RNAi identified SIN3A as a negative regulator of Insr mRNA (and JUND, MAX, and MXI as positive regulators of Irs2 mRNA). Together, our analysis identifies mechanisms which may explain the cyclic processes underlying hyperinsulinemia-induced insulin resistance in muscle, a process directly relevant to the etiology and disease progression of type 2 diabetes.

Molecular Transducers of Human Skeletal Muscle Remodeling under Different Loading States

Loading of skeletal muscle changes the tissue phenotype reflecting altered metabolic and functional demands. In humans, heterogeneous adaptation to loading complicates the identification of the underpinning molecular regulators. A within-person differential loading and analysis strategy reduces heterogeneity for changes in muscle mass by ∼40% and uses a genome-wide transcriptome method that models each mRNA from coding exons and 3' and 5' untranslated regions (UTRs). Our strategy detects ∼3-4 times more regulated genes than similarly sized studies, including substantial UTR-selective regulation undetected by other methods. We discover a core of 141 genes correlated to muscle growth, which we validate from newly analyzed independent samples (n = 100). Further validating these identified genes via RNAi in primary muscle cells, we demonstrate that members of the core genes were regulators of protein synthesis. Using proteome-constrained networks and pathway analysis reveals notable relationships with the molecular characteristics of human muscle aging and insulin sensitivity, as well as potential drug therapies.

Upstream analysis of alternative splicing: a review of computational approaches to predict context-dependent splicing factors

Alternative splicing (AS) has shown to play a pivotal role in the development of diseases, including cancer. Specifically, all the hallmarks of cancer (angiogenesis, cell immortality, avoiding immune system response, etc.) are found to have a counterpart in aberrant splicing of key genes. Identifying the context-specific regulators of splicing provides valuable information to find new biomarkers, as well as to define alternative therapeutic strategies. The computational models to identify these regulators are not trivial and require three conceptual steps: the detection of AS events, the identification of splicing factors that potentially regulate these events and the contextualization of these pieces of information for a specific experiment. In this work, we review the different algorithmic methodologies developed for each of these tasks. Main weaknesses and strengths of the different steps of the pipeline are discussed. Finally, a case study is detailed to help the reader be aware of the potential and limitations of this computational approach.

Longevity-related molecular pathways are subject to midlife "switch" in humans

Emerging evidence indicates that molecular aging may follow nonlinear or discontinuous trajectories. Whether this occurs in human neuromuscular tissue, particularly for the noncoding transcriptome, and independent of metabolic and aerobic capacities, is unknown. Applying our novel RNA method to quantify tissue coding and long noncoding RNA (lncRNA), we identified ~800 transcripts tracking with age up to ~60 years in human muscle and brain. In silico analysis demonstrated that this temporary linear "signature" was regulated by drugs, which reduce mortality or extend life span in model organisms, including 24 inhibitors of the IGF-1/PI3K/mTOR pathway that mimicked, and 5 activators that opposed, the signature. We profiled Rapamycin in nondividing primary human myotubes (n = 32 HTA 2.0 arrays) and determined the transcript signature for reactive oxygen species in neurons, confirming that our age signature was largely regulated in the "pro-longevity" direction. Quantitative network modeling demonstrated that age-regulated ncRNA equaled the contribution of protein-coding RNA within structures, but tended to have a lower heritability, implying lncRNA may better reflect environmental influences. Genes ECSIT, UNC13, and SKAP2 contributed to a network that did not respond to Rapamycin, and was associated with "neuron apoptotic processes" in protein-protein interaction analysis (FDR = 2.4%). ECSIT links inflammation with the continued age-related downwards trajectory of mitochondrial complex I gene expression (FDR < 0.01%), implying that sustained inhibition of ECSIT may be maladaptive. The present observations link, for the first time, model organism longevity programs with the endogenous but temporary genome-wide responses to aging in humans, revealing a pattern that may ultimately underpin personalized rates of health span.

Alternative splicing of the vitamin D receptor modulates target gene expression and promotes ligand-independent functions

Alternative splicing modulates gene function by creating splice variants with alternate functions or non-coding RNA activity. Naturally occurring variants of nuclear receptor (NR) genes with dominant negative or gain-of-function phenotypes have been documented, but their cellular roles, regulation, and responsiveness to environmental stress or disease remain unevaluated. Informed by observations that class I androgen and estrogen receptor variants display ligand-independent signaling in human cancer tissues, we questioned whether the function of class II NRs, like the vitamin D receptor (VDR), would also respond to alternative splicing regulation. Artificial VDR constructs lacking exon 3 (Dex3-VDR), encoding part of the DNA binding domain (DBD), and exon 8 (Dex8-VDR), encoding part of the ligand binding domain (LBD), were transiently transfected into DU-145 cells and stably-integrated into Caco-2 cells to study their effect on gene expression and cell viability. Changes in VDR promoter signaling were monitored by the expression of target genes (e.g. CYP24A1, CYP3A4 and CYP3A5). Ligand-independent VDR signaling was observed in variants lacking exon 8, and a significant loss of gene suppressor function was documented for variants lacking exon 3. The gain-of-function behavior of the Dex8-VDR variant was recapitulated in vitro using antisense oligonucleotides (ASO) that induce the skipping of exon 8 in wild-type VDR. ASO targeting the splice acceptor site of exon 8 significantly stimulated ligand-independent VDR reporter activity and the induction of CYP24A1 above controls. These results demonstrate how alternative splicing can re-program NR gene function, highlighting novel mechanisms of toxicity and new opportunities for the use of splice-switching oligonucleotides (SSO) in precision medicine.

Comparison of RNA-seq and microarray platforms for splice event detection using a cross-platform algorithm

Background

RNA-seq is a reference technology for determining alternative splicing at genome-wide level. Exon arrays remain widely used for the analysis of gene expression, but show poor validation rate with regard to splicing events. Commercial arrays that include probes within exon junctions have been developed in order to overcome this problem.

We compare the performance of RNA-seq (Illumina HiSeq) and junction arrays (Affymetrix Human Transcriptome array) for the analysis of transcript splicing events. Three different breast cancer cell lines were treated with CX-4945, a drug that severely affects splicing. To enable a direct comparison of the two platforms, we adapted EventPointer, an algorithm that detects and labels alternative splicing events using junction arrays, to work also on RNA-seq data. Common results and discrepancies between the technologies were validated and/or resolved by over 200 PCR experiments.

Results

As might be expected, RNA-seq appears superior in cases where the technologies disagree and is able to discover novel splicing events beyond the limitations of physical probe-sets. We observe a high degree of coherence between the two technologies, however, with correlation of EventPointer results over 0.90. Through decimation, the detection power of the junction arrays is equivalent to RNA-seq with up to 60 million reads.

Conclusions

Our results suggest, therefore, that exon-junction arrays are a viable alternative to RNA-seq for detection of alternative splicing events when focusing on well-described transcriptional regions.

Electronic supplementary material

The online version of this article (10.1186/s12864-018-5082-2) contains supplementary material, which is available to authorized users.

A coding and non-coding transcriptomic perspective on the genomics of human metabolic disease

Genome-wide association studies (GWAS), relying on hundreds of thousands of individuals, have revealed >200 genomic loci linked to metabolic disease (MD). Loss of insulin sensitivity (IS) is a key component of MD and we hypothesized that discovery of a robust IS transcriptome would help reveal the underlying genomic structure of MD. Using 1,012 human skeletal muscle samples, detailed physiology and a tissue-optimized approach for the quantification of coding (>18,000) and non-coding (>15,000) RNA (ncRNA), we identified 332 fasting IS-related genes (CORE-IS). Over 200 had a proven role in the biochemistry of insulin and/or metabolism or were located at GWAS MD loci. Over 50% of the CORE-IS genes responded to clinical treatment; 16 quantitatively tracking changes in IS across four independent studies (P = 0.0000053: negatively: AGL, G0S2, KPNA2, PGM2, RND3 and TSPAN9 and positively: ALDH6A1, DHTKD1, ECHDC3, MCCC1, OARD1, PCYT2, PRRX1, SGCG, SLC43A1 and SMIM8). A network of ncRNA positively related to IS and interacted with RNA coding for viral response proteins (P < 1 × 10-48), while reduced amino acid catabolic gene expression occurred without a change in expression of oxidative-phosphorylation genes. We illustrate that combining in-depth physiological phenotyping with robust RNA profiling methods, identifies molecular networks which are highly consistent with the genetics and biochemistry of human metabolic disease.

A Systems-Level Analysis Reveals Circadian Regulation of Splicing in Colorectal Cancer

Accumulating evidence points to a significant role of the circadian clock in the regulation of splicing in various organisms, including mammals. Both dysregulated circadian rhythms and aberrant pre-mRNA splicing are frequently implicated in human disease, in particular in cancer. To investigate the role of the circadian clock in the regulation of splicing in a cancer progression context at the systems-level, we conducted a genome-wide analysis and compared the rhythmic transcriptional profiles of colon carcinoma cell lines SW480 and SW620, derived from primary and metastatic sites of the same patient, respectively. We identified spliceosome components and splicing factors with cell-specific circadian expression patterns including SRSF1, HNRNPLL, ESRP1, and RBM 8A, as well as altered alternative splicing events and circadian alternative splicing patterns of output genes (e.g., VEGFA, NCAM1, FGFR2, CD44) in our cellular model. Our data reveals a remarkable interplay between the circadian clock and pre-mRNA splicing with putative consequences in tumor progression and metastasis.

The Usage of Exon-Exon Splice Junctions for the Detection of Alternative Splicing using the REIDS model

Alternative gene splicing is a common phenomenon in which a single gene gives rise to multiple transcript isoforms. The process is strictly guided and involves a multitude of proteins and regulatory complexes. Unfortunately, aberrant splicing events have been linked to genetic disorders. Therefore, understanding mechanisms of alternative splicing regulation and differences in splicing events between diseased and healthy tissues is crucial in advancing personalized medicine and drug developments. We propose a linear mixed model, Random Effects for the Identification of Differential Splicing (REIDS), for the identification of alternative splicing events using Human Transcriptome Arrays (HTA). For each exon, a splicing score is calculated based on two scores, an exon score and an array score. The junction information is used to rank the identified exons from strongly confident to less confident candidates for alternative splicing. The design of junctions was also discussed to highlight the complexity of exon-exon and exon-junction interactions. Based on a list of Rt-PCR validated probe sets, REIDS outperforms AltAnalyze and iGems in the % recall rate.

A random effects model for the identification of differential splicing (REIDS) using exon and HTA arrays

Background

Alternative gene splicing is a common phenomenon in which a single gene gives rise to multiple transcript isoforms. The process is strictly guided and involves a multitude of proteins and regulatory complexes. Unfortunately, aberrant splicing events do occur which have been linked to genetic disorders, such as several types of cancer and neurodegenerative diseases (Fan et al., Theor Biol Med Model 3:19, 2006). Therefore, understanding the mechanism of alternative splicing and identifying the difference in splicing events between diseased and healthy tissue is crucial in biomedical research with the potential of applications in personalized medicine as well as in drug development.

Results

We propose a linear mixed model, Random Effects for the Identification of Differential Splicing (REIDS), for the identification of alternative splicing events. Based on a set of scores, an exon score and an array score, a decision regarding alternative splicing can be made. The model enables the ability to distinguish a differential expressed gene from a differential spliced exon. The proposed model was applied to three case studies concerning both exon and HTA arrays.

Conclusion

The REIDS model provides a work flow for the identification of alternative splicing events relying on the established linear mixed model. The model can be applied to different types of arrays.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-017-1687-8) contains supplementary material, which is available to authorized users.

shinyheatmap: Ultra fast low memory heatmap web interface for big data genomics

BACKGROUND

Transcriptomics, metabolomics, metagenomics, and other various next-generation sequencing (-omics) fields are known for their production of large datasets, especially across single-cell sequencing studies. Visualizing such big data has posed technical challenges in biology, both in terms of available computational resources as well as programming acumen. Since heatmaps are used to depict high-dimensional numerical data as a colored grid of cells, efficiency and speed have often proven to be critical considerations in the process of successfully converting data into graphics. For example, rendering interactive heatmaps from large input datasets (e.g., 100k+ rows) has been computationally infeasible on both desktop computers and web browsers. In addition to memory requirements, programming skills and knowledge have frequently been barriers-to-entry for creating highly customizable heatmaps.

RESULTS

We propose shinyheatmap: an advanced user-friendly heatmap software suite capable of efficiently creating highly customizable static and interactive biological heatmaps in a web browser. shinyheatmap is a low memory footprint program, making it particularly well-suited for the interactive visualization of extremely large datasets that cannot typically be computed in-memory due to size restrictions. Also, shinyheatmap features a built-in high performance web plug-in, fastheatmap, for rapidly plotting interactive heatmaps of datasets as large as 105-107 rows within seconds, effectively shattering previous performance benchmarks of heatmap rendering speed.

CONCLUSIONS

shinyheatmap is hosted online as a freely available web server with an intuitive graphical user interface: http://shinyheatmap.com. The methods are implemented in R, and are available as part of the shinyheatmap project at: https://github.com/Bohdan-Khomtchouk/shinyheatmap. Users can access fastheatmap directly from within the shinyheatmap web interface, and all source code has been made publicly available on Github: https://github.com/Bohdan-Khomtchouk/fastheatmap.

Molecular Diagnostics of Ageing and Tackling Age-related Disease

As average life expectancy increases there is a greater focus on health-span and, in particular, how to treat or prevent chronic age-associated diseases. Therapies which were able to control 'biological age' with the aim of postponing chronic and costly diseases of old age require an entirely new approach to drug development. Molecular technologies and machine-learning methods have already yielded diagnostics that help guide cancer treatment and cardiovascular procedures. Discovery of valid and clinically informative diagnostics of human biological age (combined with disease-specific biomarkers) has the potential to alter current drug-discovery strategies, aid clinical trial recruitment and maximize healthy ageing. I will review some basic principles that govern the development of 'ageing' diagnostics, how such assays could be used during the drug-discovery or development process. Important logistical and statistical considerations are illustrated by reviewing recent biomarker activity in the field of Alzheimer's disease, as dementia represents the most pressing of priorities for the pharmaceutical industry, as well as the chronic disease in humans most associated with age.

Biomarkers of browning of white adipose tissue and their regulation during exercise- and diet-induced weight loss

BACKGROUND

A hypothesis exists whereby an exercise- or dietary-induced negative energy balance reduces human subcutaneous white adipose tissue (scWAT) mass through the formation of brown-like adipocyte (brite) cells. However, the validity of biomarkers of brite formation has not been robustly evaluated in humans, and clinical data that link brite formation and weight loss are sparse.

OBJECTIVES

We used rosiglitazone and primary adipocytes to stringently evaluate a set of biomarkers for brite formation and determined whether the expression of biomarker genes in scWAT could explain the change in body composition in response to exercise training combined with calorie restriction in obese and overweight women (n = 79).

DESIGN

Gene expression was derived from exon DNA microarrays and preadipocytes from obesity-resistant and -sensitive mice treated with rosiglitazone to generate candidate brite biomarkers from a microarray. These biomarkers were evaluated against data derived from scWAT RNA from obese and overweight women before and after supervised exercise 5 d/wk for 16 wk combined with modest calorie restriction (∼0.84 MJ/d).

RESULTS

Forty percent of commonly used brite gene biomarkers exhibited an exon or strain-specific regulation. No biomarkers were positively related to weight loss in human scWAT. Greater weight loss was significantly associated with less uncoupling protein 1 expression (P = 0.006, R(2) = 0.09). In a follow-up global analysis, there were 161 genes that covaried with weight loss that were linked to greater CCAAT/enhancer binding protein α activity (z = 2.0, P = 6.6 × 10(-7)), liver X receptor α/β agonism (z = 2.1, P = 2.8 × 10(-7)), and inhibition of leptin-like signaling (z = -2.6, P = 3.9 × 10(-5)).

CONCLUSION

We identify a subset of robust RNA biomarkers for brite formation and show that calorie-restriction-mediated weight loss in women dynamically remodels scWAT to take on a more-white rather than a more-brown adipocyte phenotype.