NCI Cancer Research Data Commons: Lessons Learned and Future State

More than ever, scientific progress in cancer research hinges on our ability to combine datasets and extract meaningful interpretations to better understand diseases and ultimately inform the development of better treatments and diagnostic tools. To enable the successful sharing and use of big data, the NCI developed the Cancer Research Data Commons (CRDC), providing access to a large, comprehensive, and expanding collection of cancer data. The CRDC is a cloud-based data science infrastructure that eliminates the need for researchers to download and store large-scale datasets by allowing them to perform analysis where data reside. Over the past 10 years, the CRDC has made significant progress in providing access to data and tools along with training and outreach to support the cancer research community. In this review, we provide an overview of the history and the impact of the CRDC to date, lessons learned, and future plans to further promote data sharing, accessibility, interoperability, and reuse. See related articles by Brady et al., p. 1384, Wang et al., p. 1388, and Pot et al., p. 1396.

Pan-African genome demonstrates how population-specific genome graphs improve high-throughput sequencing data analysis

Graph-based genome reference representations have seen significant development, motivated by the inadequacy of the current human genome reference to represent the diverse genetic information from different human populations and its inability to maintain the same level of accuracy for non-European ancestries. While there have been many efforts to develop computationally efficient graph-based toolkits for NGS read alignment and variant calling, methods to curate genomic variants and subsequently construct genome graphs remain an understudied problem that inevitably determines the effectiveness of the overall bioinformatics pipeline. In this study, we discuss obstacles encountered during graph construction and propose methods for sample selection based on population diversity, graph augmentation with structural variants and resolution of graph reference ambiguity caused by information overload. Moreover, we present the case for iteratively augmenting tailored genome graphs for targeted populations and demonstrate this approach on the whole-genome samples of African ancestry. Our results show that population-specific graphs, as more representative alternatives to linear or generic graph references, can achieve significantly lower read mapping errors and enhanced variant calling sensitivity, in addition to providing the improvements of joint variant calling without the need of computationally intensive post-processing steps.

Pan-cancer analysis of whole genomes

Cancer is driven by genetic change, and the advent of massively parallel sequencing has enabled systematic documentation of this variation at the whole-genome scale1-3. Here we report the integrative analysis of 2,658 whole-cancer genomes and their matching normal tissues across 38 tumour types from the Pan-Cancer Analysis of Whole Genomes (PCAWG) Consortium of the International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA). We describe the generation of the PCAWG resource, facilitated by international data sharing using compute clouds. On average, cancer genomes contained 4-5 driver mutations when combining coding and non-coding genomic elements; however, in around 5% of cases no drivers were identified, suggesting that cancer driver discovery is not yet complete. Chromothripsis, in which many clustered structural variants arise in a single catastrophic event, is frequently an early event in tumour evolution; in acral melanoma, for example, these events precede most somatic point mutations and affect several cancer-associated genes simultaneously. Cancers with abnormal telomere maintenance often originate from tissues with low replicative activity and show several mechanisms of preventing telomere attrition to critical levels. Common and rare germline variants affect patterns of somatic mutation, including point mutations, structural variants and somatic retrotransposition. A collection of papers from the PCAWG Consortium describes non-coding mutations that drive cancer beyond those in the TERT promoter4; identifies new signatures of mutational processes that cause base substitutions, small insertions and deletions and structural variation5,6; analyses timings and patterns of tumour evolution7; describes the diverse transcriptional consequences of somatic mutation on splicing, expression levels, fusion genes and promoter activity8,9; and evaluates a range of more-specialized features of cancer genomes8,10-18.

ALS-implicated protein TDP-43 sustains levels of STMN2, a mediator of motor neuron growth and repair

The findings that amyotrophic lateral sclerosis (ALS) patients almost universally display pathological mislocalization of the RNA-binding protein TDP-43 and that mutations in its gene cause familial ALS have nominated altered RNA metabolism as a disease mechanism. However, the RNAs regulated by TDP-43 in motor neurons and their connection to neuropathy remain to be identified. Here we report transcripts whose abundances in human motor neurons are sensitive to TDP-43 depletion. Notably, expression of STMN2, which encodes a microtubule regulator, declined after TDP-43 knockdown and TDP-43 mislocalization as well as in patient-specific motor neurons and postmortem patient spinal cord. STMN2 loss upon reduced TDP-43 function was due to altered splicing, which is functionally important, as we show STMN2 is necessary for normal axonal outgrowth and regeneration. Notably, post-translational stabilization of STMN2 rescued neurite outgrowth and axon regeneration deficits induced by TDP-43 depletion. We propose that restoring STMN2 expression warrants examination as a therapeutic strategy for ALS.

Comparative genomic analysis of embryonic, lineage-converted and stem cell-derived motor neurons

Advances in stem cell science allow the production of different cell types in vitro either through the recapitulation of developmental processes, often termed 'directed differentiation', or the forced expression of lineage-specific transcription factors. Although cells produced by both approaches are increasingly used in translational applications, their quantitative similarity to their primary counterparts remains largely unresolved. To investigate the similarity between in vitro-derived and primary cell types, we harvested and purified mouse spinal motor neurons and compared them with motor neurons produced by transcription factor-mediated lineage conversion of fibroblasts or directed differentiation of pluripotent stem cells. To enable unbiased analysis of these motor neuron types and their cells of origin, we then subjected them to whole transcriptome and DNA methylome analysis by RNA sequencing (RNA-seq) and reduced representation bisulfite sequencing (RRBS). Despite major differences in methodology, lineage conversion and directed differentiation both produce cells that closely approximate the primary motor neuron state. However, we identify differences in Fas signaling, the Hox code and synaptic gene expression between lineage-converted and directed differentiation motor neurons that affect their utility in translational studies.

Using the Seven Bridges Cancer Genomics Cloud to Access and Analyze Petabytes of Cancer Data

Next-generation sequencing has produced petabytes of data, but accessing and analyzing these data remain challenging. Traditionally, researchers investigating public datasets like The Cancer Genome Atlas (TCGA) would download the data to a high-performance cluster, which could take several weeks even with a highly optimized network connection. The National Cancer Institute (NCI) initiated the Cancer Genomics Cloud Pilots program to provide researchers with the resources to process data with cloud computational resources. We present protocols using one of these Cloud Pilots, the Seven Bridges Cancer Genomics Cloud (CGC), to find and query public datasets, bring your own data to the CGC, analyze data using standard or custom workflows, and benchmark tools for accuracy with interactive analysis features. These protocols demonstrate that the CGC is a data-analysis ecosystem that fully empowers researchers with a variety of areas of expertise and interests to collaborate in the analysis of petabytes of data. © 2017 by John Wiley & Sons, Inc.

Abstract 2595: Enabling petabyte-scale cancer genomics with the NCI Cancer Cloud Pilots

The advent of next generation sequencing has accelerated the generation of genomic data and created a need for methodologies to organize, share, and analyze large volumes of data. To date, petabytes of multi-dimensional information from thousands of patients have been collected. Access and analysis of this information becomes increasingly challenging as the amount of data grows. This difficulty is exemplified when we consider data generated by the efforts of The Cancer Genomics Atlas (TCGA) network, which encompasses more than 2.5 petabytes. Historically, downloading the complete TCGA repository can require several weeks with a highly optimized network connection and access to large institutional compute clusters to perform integrated analysis, which is out of reach for many researchers. The Cancer Cloud Pilot project seeks to directly address these challenges by co-localizing data with the computational resources to analyze it where researchers can access it securely and easily. The project was born out of the recognition that conducting biological research is increasingly computationally-intensive and new approaches are required to support effective data discovery, storage, computation, and collaboration. Funded by the National Cancer Institute, the Cancer Genomics Cloud (CGC) enables researchers to leverage the power of cloud computing to gain actionable insights about cancer biology and human genetics from massive public datasets including TCGA and the Cancer Cell Line Encyclopedia on the CGC. Our approach to create a cancer cloud platform includes collaborative tools, security permissions, data harmonization, and making the data easier to query through the use of metadata curation, resource description frameworks, and visual tools. Additionally, we implemented the Common Workflow Language, an emerging standard for describing computational workflows, to support computational reproducibility. To date, more than 1200 researchers have accessed and analyzed TCGA and analyzed more than 50000 samples on the CGC since its launch in February 2016. In addition to the motivation, inception, and development of the CGC, we will present a case study on the application of unsupervised learning methods to identify individual cell types within tumors using RNA Sequencing data from TCGA cohorts. We will demonstrate how these computationally-intensive methods are benefited by the cloud and how researchers can apply open pipelines to interrogate cancer subtypes and mixed cell populations from TCGA data on their own data.

Citation Format: Gaurav Kaushik, Yilong Li, Erik Lehnert, Zeynep Onder, Devin Locke, Brandi N. Davis-Dusenbery, Deniz Kural. Enabling petabyte-scale cancer genomics with the NCI Cancer Cloud Pilots [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 2595. doi:10.1158/1538-7445.AM2017-2595

Reproducible, Scalable Fusion Gene Detection from RNA-Seq

Chromosomal rearrangements resulting in the creation of novel gene products, termed fusion genes, have been identified as driving events in the development of multiple types of cancer. As these gene products typically do not exist in normal cells, they represent valuable prognostic and therapeutic targets. Advances in next-generation sequencing and computational approaches have greatly improved our ability to detect and identify fusion genes. Nevertheless, these approaches require significant computational resources. Here we describe an approach which leverages cloud computing technologies to perform fusion gene detection from RNA sequencing data at any scale. We additionally highlight methods to enhance reproducibility of bioinformatics analyses which may be applied to any next-generation sequencing experiment.

Genetic validation of a therapeutic target in a mouse model of ALS

Neurons produced from stem cells have emerged as a tool to identify new therapeutic targets for neurological diseases such as amyotrophic lateral sclerosis (ALS). However, it remains unclear to what extent these new mechanistic insights will translate to animal models, an important step in the validation of new targets. Previously, we found that glia from mice carrying the SOD1G93A mutation, a model of ALS, were toxic to stem cell-derived human motor neurons. We use pharmacological and genetic approaches to demonstrate that the prostanoid receptor DP1 mediates this glial toxicity. Furthermore, we validate the importance of this mechanism for neural degeneration in vivo. Genetic ablation of DP1 in SOD1G93A mice extended life span, decreased microglial activation, and reduced motor neuron loss. Our findings suggest that blocking DP1 may be a therapeutic strategy in ALS and demonstrate that discoveries from stem cell models of disease can be corroborated in vivo.

Pathways disrupted in human ALS motor neurons identified through genetic correction of mutant SOD1

Although many distinct mutations in a variety of genes are known to cause Amyotrophic Lateral Sclerosis (ALS), it remains poorly understood how they selectively impact motor neuron biology and whether they converge on common pathways to cause neuronal degeneration. Here, we have combined reprogramming and stem cell differentiation approaches with genome engineering and RNA sequencing to define the transcriptional and functional changes that are induced in human motor neurons by mutant SOD1. Mutant SOD1 protein induced a transcriptional signature indicative of increased oxidative stress, reduced mitochondrial function, altered subcellular transport, and activation of the ER stress and unfolded protein response pathways. Functional studies demonstrated that these pathways were perturbed in a manner dependent on the SOD1 mutation. Finally, interrogation of stem-cell-derived motor neurons produced from ALS patients harboring a repeat expansion in C9orf72 indicates that at least a subset of these changes are more broadly conserved in ALS.

How to make spinal motor neurons

All muscle movements, including breathing, walking, and fine motor skills rely on the function of the spinal motor neuron to transmit signals from the brain to individual muscle groups. Loss of spinal motor neuron function underlies several neurological disorders for which treatment has been hampered by the inability to obtain sufficient quantities of primary motor neurons to perform mechanistic studies or drug screens. Progress towards overcoming this challenge has been achieved through the synthesis of developmental biology paradigms and advances in stem cell and reprogramming technology, which allow the production of motor neurons in vitro. In this Primer, we discuss how the logic of spinal motor neuron development has been applied to allow generation of motor neurons either from pluripotent stem cells by directed differentiation and transcriptional programming, or from somatic cells by direct lineage conversion. Finally, we discuss methods to evaluate the molecular and functional properties of motor neurons generated through each of these techniques.

Nanog-independent reprogramming to iPSCs with canonical factors

It has been suggested that the transcription factor Nanog is essential for the establishment of pluripotency during the derivation of embryonic stem cells and induced pluripotent stem cells (iPSCs). However, successful reprogramming to pluripotency with a growing list of divergent transcription factors, at ever-increasing efficiencies, suggests that there may be many distinct routes to a pluripotent state. Here, we have investigated whether Nanog is necessary for reprogramming murine fibroblasts under highly efficient conditions using the canonical-reprogramming factors Oct4, Sox2, Klf4, and cMyc. In agreement with prior results, the efficiency of reprogramming Nanog (-/-) fibroblasts was significantly lower than that of control fibroblasts. However, in contrast to previous findings, we were able to reproducibly generate iPSCs from Nanog (-/-) fibroblasts that effectively contributed to the germline of chimeric mice. Thus, whereas Nanog may be an important mediator of reprogramming, it is not required for establishing pluripotency in the mouse, even under standard conditions.

Atrial natriuretic peptide is negatively regulated by microRNA-425

Numerous common genetic variants have been linked to blood pressure, but no underlying mechanism has been elucidated. Population studies have revealed that the variant rs5068 (A/G) in the 3' untranslated region of NPPA, the gene encoding atrial natriuretic peptide (ANP), is associated with blood pressure. We selected individuals on the basis of rs5068 genotype (AG vs. AA) and fed them a low- or high-salt diet for 1 week, after which they were challenged with an intravenous saline infusion. On both diets, before and after saline administration, ANP levels were up to 50% higher in AG individuals than in AA individuals, a difference comparable to the changes induced by high-salt diet or saline infusion. In contrast, B-type natriuretic peptide levels did not differ by rs5068 genotype. We identified a microRNA, miR-425, that is expressed in human atria and ventricles and is predicted to bind the sequence spanning rs5068 for the A, but not the G, allele. miR-425 silenced NPPA mRNA in an allele-specific manner, with the G allele conferring resistance to miR-425. This study identifies miR-425 as a regulator of ANP production, raising the possibility that miR-425 antagonists could be used to treat disorders of salt overload, including hypertension and heart failure.

Inhibition of microRNA-302 (miR-302) by bone morphogenetic protein 4 (BMP4) facilitates the BMP signaling pathway

The signaling pathway mediated by BMPs plays an essential role during development as well as the maintenance of homeostasis in adult. Aberrant activation or inactivation of BMP signaling can lead to developmental defects and various human disorders. To fine-tune its activity, BMP signaling is regulated both positively and negatively by extrinsic and intrinsic regulatory factors that modulate binding of ligand to the receptors, and the activity of receptors and their dedicated signal transducers, the Smad proteins. Upon BMP binding to the receptor complex, Smad proteins translocate to the nucleus and modulate gene expression transcriptionally by directly associating with the promoter region of target genes, or post-transcriptionally through modulation of microRNA (miRNA) synthesis. In this study, we demonstrate that BMP signaling down-regulates transcription of the miRNA-302∼367 gene cluster. We show that the type II BMP receptor (BMPRII) is a novel target of miR-302. Upon overexpression, miR-302 targets a partially complementary sequence localized in the 3'-untranslated region (UTR) of BMPRII transcripts and leads to destabilization of the transcripts and down-regulation of BMP signaling. We propose that the negative regulatory loop of BMP4-miR-302-BMPRII is a potential mechanism for the maintenance and fine-tuning of the BMP signaling pathway in various systems.

MicroRNA in Cancer: The Involvement of Aberrant MicroRNA Biogenesis Regulatory Pathways

MicroRNAs (miRNAs) are small, noncoding RNAs that influence diverse biological outcomes through the repression of target genes during normal development and pathological responses. In particular, the alteration of miRNA expression has dramatic consequences for the progression of tumorigenesis. miRNAs undergo two processing steps that transform a long primary transcript into the mature miRNA. Although the general miRNA biogenesis pathway is well established, it is clear that not all miRNAs are created equally. Recent studies show that miRNA expression is controlled by diverse mechanisms in response to cellular stimuli. In this review, we discuss the mechanisms that govern the regulation of miRNA biogenesis with particular focus on how these mechanisms are perturbed in cancer.

Bone morphogenetic protein signaling in vascular disease: anti-inflammatory action through myocardin-related transcription factor A

Pulmonary artery hypertension (PAH) patients exhibit elevated levels of inflammatory cytokines and infiltration of inflammatory cells in the lung. Concurrently, mutations of bmpr2, the gene encoding the type II receptor of bone morphogenetic proteins (BMP), are found in ∼75% of patients with familial PAH, but a possible nexus between increased inflammation and diminished BMP signaling has hitherto remained elusive. We previously showed that BMP4 triggers nuclear localization of the Myocardin-related transcription factor A (MRTF-A) in human pulmonary artery smooth muscle cells (PASMC), resulting in the induction of contractile proteins. Here we report the BMPR2-dependent repression of a set of inflammatory mediators in response to BMP4 stimulation of PASMC. Forced expression of MRTF-A precisely emulates the anti-inflammatory effect of BMP4, while MRTF-A depletion precludes BMP4-mediated cytokine inhibition. BMP4 and MRTF-A block signaling through NF-κB, the keystone of most pathways leading to inflammatory responses, at the level of chromatin recruitment and promoter activation. Moreover, MRTF-A physically interacts with RelA/p65, the NF-κB subunit endowed with a transcription activation domain. Interestingly, the MRTF-A-NF-κB interaction is mutually antagonistic: stimulation of NF-κB signaling by TNFα, as well as p65 overexpression, hinders MRTF-A activity and the expression of contractile genes. Thus, a molecular inhibitory pathway linking BMP4 signaling, activation of MRTF-A, and inhibition of NF-κB provides insights into the etiology of PAH and a potential focus of therapeutic intervention.

Micromanaging vascular smooth muscle cell differentiation and phenotypic modulation

The phenotype of vascular smooth muscle cells (VSMCs) is dynamically regulated in response to various stimuli. In a cellular process known as phenotype switching, VSMCs alternate between a contractile and synthetic phenotype state. Deregulation of phenotype switching is associated with vascular disorders such as atherosclerosis, restenosis after angioplasty, and pulmonary hypertension. An important role for microRNAs (miRNAs) in VSMC development and phenotype switching has recently been uncovered. Individual miRNAs are involved in promoting both contractile and synthetic VSMC phenotype. In this review, we summarize recent advances in the understanding of miRNA function in the regulation of VSMC phenotype regulation.

Bone morphogenetic protein 4 promotes vascular smooth muscle contractility by activating microRNA-21 (miR-21), which down-regulates expression of family of dedicator of cytokinesis (DOCK) proteins

The bone morphogenetic protein 4 (BMP4) signaling pathway plays a critical role in the promotion and maintenance of the contractile phenotype in vascular smooth muscle cell (vSMC). Misexpression or inactivating mutations of the BMP receptor gene can lead to dedifferentiation of vSMC characterized by increased migration and proliferation that is linked to vascular proliferative disorders. Previously we demonstrated that vSMCs increase microRNA-21 (miR-21) biogenesis upon BMP4 treatment, which induces contractile gene expression by targeting programmed cell death 4 (PDCD4). To identify novel targets of miR-21 that are critical for induction of the contractile phenotype by BMP4, biotinylated miR-21 was expressed in vSMCs followed by an affinity purification of mRNAs associated with miR-21. Nearly all members of the dedicator of cytokinesis (DOCK) 180-related protein superfamily were identified as targets of miR-21. Down-regulation of DOCK4, -5, and -7 by miR-21 inhibited cell migration and promoted cytoskeletal organization by modulating an activity of small GTPase. Thus, this study uncovers a regulatory mechanism of the vSMC phenotype by the BMP4-miR-21 axis through DOCK family proteins.

down-regulation of Kruppel-like factor-4 (KLF4) by microRNA-143/145 is critical for modulation of vascular smooth muscle cell phenotype by transforming growth factor-beta and bone morphogenetic protein 4

In the postnatal vasculature, fully differentiated and quiescent vascular smooth muscle cells (VSMCs) in a "contractile" phenotype are required for the normal regulation of vascular tone. The transforming growth factor-β (TGF-β) superfamily of growth factors (TGF-βs and bone morphogenetic proteins (BMPs)) are potent inducers of contractile phenotype and mediate (i) induction of contractile genes, and (ii) inhibition of VSMC growth and migration. Transcription of contractile genes is positively regulated by a regulatory DNA element called a CArG box. The CArG box is activated by the binding of serum response factor and its coactivators, myocardin (Myocd) or Myocd-related transcription factors (MRTFs). Krüppel-like factor-4 (KLF4) is known to inhibit activation of the CArG box. However, the potential role of KLF4 in the contractile activities of TGF-β or BMP has not been explored. Here, we demonstrate that TGF-β and BMP4 rapidly down-regulate KLF4 through induction of microRNA-143 (miR-143) and miR-145, which leads to a reduction of KLF4 transcripts and decreased KLF4 protein expression. Inhibition of miR-145 prevents down-regulation of KLF4 and activation of contractile genes by TGF-β or BMP4, suggesting that modulation of KLF4 is a prerequisite for induction of contractile genes by TGF-β and BMP4. Interestingly, both TGF-β and BMP4 activate transcription of the miR-143/145 gene cluster through the CArG box, however, TGF-β mediates this effect through induction of Myocd expression, whereas BMP4 utilizes nuclear translocation of MRTF-A. Thus, this study sheds light on both the similarities and the differences of TGF-β and BMP4 signaling in the regulation of KLF4 and contractile genes.

Smad-mediated miRNA processing: a critical role for a conserved RNA sequence

microRNAs (miRNAs) are short, 21-24 nucleotide (nt), non-coding RNAs that post-transcriptionally regulate the expression of messenger RNAs (mRNAs). Through the regulation of their cognate mRNAs, miRNAs control diverse aspects of biology, including development, cellular differentiation, proliferation, metabolism, and death. Thus, miRNAs play a critical role in the determination of normal cellular physiology and misexpression of miRNAs leads to pathological responses. Understanding the mechanisms that control miRNA expression is an important step forward as novel functions of miRNAs continue to be uncovered. In addition to transcriptional regulation, multiple pathways of post-transcriptional modulation of miRNA expression have been uncovered. In this review we discuss the role of the Smads in the regulation of miRNA processing in response to Transforming Growth Factor-β stimulation.

Mechanisms of control of microRNA biogenesis

MicroRNAs (miRNAs) are a class of ∼22 nt non-coding RNAs that control diverse biological functions in animals, plants and unicellular eukaryotes by promoting degradation or inhibition of translation of target mRNAs. miRNA expression is often tissue specific and developmentally regulated. Aberrant expression of miRNAs has been linked to developmental abnormalities and human diseases, including cancer and cardiovascular disorders. The recent identification of mechanisms of miRNA biogenesis regulation uncovers that various factors or growth factor signalling pathways control every step of the miRNA biogenesis pathway. Here, we review the mechanisms that control the regulation of miRNA biogenesis discovered in human cells. Further understanding of the mechanisms that control of miRNA biogenesis may allow the development of tools to modulate the expression of specific miRNAs, which is crucial for the development of novel therapies for human disorders derived from aberrant expression of miRNAs.