Search Results - EGA European Genome-Phenome Archive

Data upcycling, powered by EGA

In 2012 a collaborative effort headed by the teams of Carlos Caldas in Cambridge, UK and Samuel Aparicio, Canada, produced a pivotal study aimed at defining the mutational landscape of breast cancer. They gathered a collection of over 2000 breast cancer samples (METABRIC) associated with long-term clinical follow-up. They delineated an array of inherited and somatically acquired genetic variants and associated them to gene expression. This led to the identification of putative cancer-driving events and improved stratification of breast cancer patients. The study elicited huge attention and was tackled as herculean effort by fellow researchers. Cristina Curtis and her colleagues deposited the data in the EGA archive (EGAS00000000083). What they probably could not foresee is that their data would be reused more than 100 times, becoming one of the most downloaded datasets at EGA. In 2015, Frederik S. Varn and colleagues accessed the above-mentioned data from the New Hampshire, USA, to correlate it to transcriptional programs of hematopoietic cells infiltrated in breast tumours and found that those can influence patients' prognosis. Their findings, published in Nature Communications, improved the understanding of the interplay between immune system and cancer. Again, in 2017 the dataset was downloaded by an Italian/Polish collaboration and used to corroborate their mechanistic study that demonstrate how a long non-coding RNA (lncRNA) is aberrantly localized into the chromatin, where it modulates oncogenic isoforms expression, thus contributing to breast cancer. To date, the METABRIC dataset has been reused to contribute to 147 publications. Its impact on breast cancer research is incalculable, and it even spread beyond that, with recent contributions to the understanding of pan-cancer mechanisms and features. Some of the publications are translating into improvements in diagnosis and treatment of different types of breast cancer. This is just an example of how data upcycling can amplify the potential of any dataset, well past the scope of its creation, the imagination of its owners and any geographical border. At EGA, we are proud to empower such fruitful worldwide cycles of knowledge, providing a platform that enables safe sharing of sensitive genetic data.

Blog data-upcycling-powered-by-ega

Gut microbiome dynamics unravelled with metagenomics sequencing

It has been already several years since we discovered that microbial therapeutics like Faecal Microbiota Transplantation can be used to treat Clostridium difficile infection (CDI) and Inflammatory bowel disease (IBD) with great success. The concept has challenged us to test the limits of our knowledge with the perspective to improve health in the most diverse conditions like Parkinson disease, depression, autism and obesity, just to mention a few. Note: the figure belongs to the original paper and is owned by the authors. Allyson L.Bird and colleagues from Genentech Inc. (San Francisco) recently published an interesting study building up on the key concept that “As microbial therapeutics are increasingly being tested in diverse patient populations, it is essential to understand the host and environmental factors influencing the microbiome.” The authors of this work analysed 1359 gut microbiome samples using shotgun metagenomics sequencing. The results of the comparison done between microbiota composition across the different groups showed that biological sex is the strongest driver of diversity, while aging (age range studied: 20–69) appears to be the second factor. As smartly outlined in the authors’ own graphical abstract (figure to the left), Bacteroidota species consistently increased with donors' age while Actinobacteriota species, including Bifidobacterium, decreased. This study also integrated data from oncological patients to draw cancer-related evidences. After correcting the biases for technical and geographic differences, they found that cancer patients show significantly altered gut bacterial communities compared with their healthy counterparts. Eg. E.coli was found to be increased in cancer patients together with Bacteroidota/Firmicutes ratios. By comparing sequential samples taken from the same donors (413) within short-distance time points, the authors found that in the short term, intra-individual differences were less than the inter-individual ones, indicating a certain community stability being the norm. One of the pivotal points from this interesting study is that the data proceeds from extremely well-defined and selected donors. Specifically, the data for 949 selected healthy donors came from the Milieu Intérieur (MI) Consortium and extensive metadata, including demographic variables, serological measures, dietary information, and systemic immune profiles, are available. This precise stratification allowed for the robust detection of smaller changes, compared to former works. All sequencing data has been deposited in the EGA archive under the accession number EGAS00001004437. Enjoy this amazing open access paper published in the Experimental Journal of Medicine on January 2021.

Blog gut-microbiome-dynamics-unravelled

Discover Cancer Image Europe, the first release of the EUCAIM platform to fuel cancer research and data sharing

The EUCAIM consortium and the European Commission have announced the release of Cancer Image Europe, a platform to fuel collaborative innovation and data sharing across Europe. This constitutes a major milestone in EUCAIM's development and an exciting step towards achieving the project's vision and goals. The platform aims to accelerate the pace of AI development and other data-intensive cancer research activities, constituting the basis for a fully-fledged infrastructure for sharing, reusing, and exploiting cancer imaging data, especially using AI techniques. A platform paving the way for the future of cancer diagnosis and treatment Cancer Image Europe brings benefits to researchers, clinicians, and AI innovators across Europe as it addresses the fragmentation of existing cancer image repositories. The platform gathers a public catalogue of cancer imaging datasets from the repositories of the EU-funded AI for Health Imaging projects. A distributed Atlas of Cancer Imaging with over 60 million anonymised cancer image data from over 100,000 patients will be established as part of future updates. The first version of Cancer Image Europe also includes a public catalogue of cancer imaging datasets, a federated searching tool to understand the information available in the federated providers, and full integration with Life Science Login. Moreover, it reuses and adds value to key components of EU-funded research projects and infrastructure in the field of cancer. The Public Catalogue When we talk about data re-use, data discovery is key. The EUCAIM public catalogue is the result of a systematic and collaborative process to which the EGA-CRG has actively contributed. This browsing interface gathers the metadata of the available cancer imaging datasets. This descriptive information will help researchers in their data retrieval. Our contribution to the Public Catalogue comes from our experience gained in the deployment of our federated network, the Federated EGA (FEGA), and as partners in projects such as EuCanImage (AI4HI initiative), in which we worked on the definition of common data models, ontologies and data standards. The EUCAIM Project EUCAIM is a cornerstone of the European Commission-initiated European Cancer Imaging Initiative, a flagship of the Europe's Beating Cancer Plan, which aims to foster innovation and deployment of digital technologies in cancer treatment and care, to achieve more precise and faster clinical decision-making, diagnostics, treatments, and predictive medicine for cancer patients. For more information about the first release of the Cancer Image Europe platform and the EUCAIM project, please visit cancerimage.eu.

Blog discover-cancer-image-europe

First datasets available in the Federated EGA Network

In February 2024, a great milestone was reached for the human omics data community: the first datasets are now available in the Federated European Genome-phenome Archive (FEGA). These datasets - hosted by national nodes in Poland, Norway, and Sweden - are discoverable in the EGA portal and accessible according to national laws. These datasets include: From FEGA Poland - The Genomic Map Of Poland From FEGA Norway - CRCbiome: Gut metagenome of Norwegian screening participants using FIT sampling Also from FEGA Norway - Single cell transcriptomics in expanded Tregs of APS-1 patients From FEGA Sweden - A Combined Omics and Tissue Biobank for Paediatric Cancers The establishment of these datasets represents a significant step towards achieving the broader objectives of the FEGA; to facilitate the sharing of high-quality human genomic and phenotypic data across borders. “The value for users is twofold: it becomes easier to make data available securely compared to storing data locally, and data becomes easier to find for other researchers. The possibility of reuse improves the chances of both verifying scientific discoveries and reaching groundbreaking conclusions, of benefit to the research community and later also patients,” said Anna Hagwall, project manager and head of human data at FEGA Sweden, NBIS. About Federated EGA FEGA was launched in 2022 as a network to store and lawfully share sensitive human omics data safely. It relies on the technology originally developed by the EGA and support from ongoing collaborations with ELIXIR - a European life sciences infrastructure. FEGA has contributed to and benefitted from the implementation of community standards for sensitive data management including those developed by the Global Alliance for Genomics and Health (GA4GH). “The FEGA concept and the development of the Nordic FEGA nodes have been made possible thanks to several projects coordinated by NeIC,” said Kjell Petersen, FEGA Norway technical lead. Managing human omics data is challenging due to its sheer size and privacy concerns, often resulting in restrictions on how and where sensitive data can be shared. The FEGA network follows a federated data access model whereby sensitive data is stored securely within a local or national jurisdiction, while discovery across the network for data of interest is enabled via a central portal. As of February 2024, the FEGA network comprises seven European national nodes with tens of additional national partners working towards establishing a FEGA node. As the FEGA network continues to onboard nodes from around the world, populated with more and more data, we move closer to a vision of a truly global resource for discovery of and access to sensitive human omics data to accelerate disease research and improve human health. "The Federated EGA network is entering a key phase where the first research datasets are now available across the network. In coming years, we look forward to expanding these datasets to cover key areas such as cancer and rare diseases for the benefit of patients" said Thomas Keane, Group Leader at EMBL-EBI.

Blog first-datasets-available-in-Federated-EGA

How mitochondrial DNA research can benefit from data reuse through EGA?

Mitochondria are the only cellular organelles harbouring their own DNA, which encodes for proteins essentials in its respiration process. Despite its discovery in 1960, Mitochondrial DNA is often overlooked compared to nuclear DNA. Indeed, not all sequencing technologies capture a good representation of mtDNA (mitochondrial DNA). Recently, Ed Reznik's team (The Ed Reznik Lab | Memorial Sloan Kettering Cancer Center) carried out a massive analysis and reanalysis of single cells WGS (whole genome sequencing) produced with an amplification-free protocol (Direct library prep, DLP). As they demonstrated in the paper Single-cell mtDNA dynamics in tumors is driven by coregulation of nuclear and mitochondrial genomes - PMC, DLP is more effective in catching mtDNA compared to classical single-cell sequencing. These data allowed the team to unravel mtDNA dynamics in tumoral and normal samples, highlighting how mtDNA and nuclear DNA are co-regulated and their ratio is critical for phenotype. Data reuse to achieve the greatest results For this project, Reznik's team re-used data from 3 studies stored at the European Genome-phenome Archive (EGAS00001006343, EGAS00001004448 and EGAS00001003190) that comprised a total of 602 datasets. All of them were initially produced by the teams of Sam Aparicio and Sohrab P. Shah at the Memorial Sloan Kettering Cancer Centre, where Ed Reznik is a promising group leader. This demonstrates, once again, that networking, collaboration and data sharing can achieve the greatest results when coupled with innovative and smart ideas. At the EGA we are proud to provide the scientific community with an infrastructure to enable all those very smart ideas out there to find the right data to be tested.

Blog mitochondrial-dna-research

Drug development for pediatric cancers: the importance of patient' s genomic data re-use

Cancer in children is uncommon and the overall prognosis for most pediatric cancers is good. However, while the combined survival rates have improved over the last decades, certain childhood malignancies, such as high-grade gliomas or metastatic sarcomas, still remain incurable in most patients. New strategies for targeting those devastating diseases are imperative, and patient genomic data can become a key asset in the process [1]. The Pediatric Cancer Genome Project’s datasets At EGA, we store over 500 datasets with sequencing data from pediatric cancer patients. A remarkable case is the datasets belonging to the Pediatric Cancer Genome Project (PCGP) from St. Jude Children’s Research Hospital–Washington University (Table 1 - EGAS list). PCGP is an ambitious effort to identify the mutations that drive childhood cancer and find new cures. Those datasets include 600 patients with complete tumor and normal genomes from 15 different tumor types. PCGP datasets have been an important resource for various studies that pooled genomic data from different public and in-house datasets to perform extensive genomic characterizations and publish comprehensive pan-cancer pediatric studies [2],[3]. Other interesting pediatric cancer datasets are from the Pediatric Brain Tumor Consortium, Sickkids or ICGC PedBrain project. How data reuse impacts drug development for pediatric cancers Recently, in October 2024, Nature Communications published a work led by the University of Michigan where authors used data from EGA (PCGP) check the table below as part of their strategy to pursue the identification of new tumor vulnerabilities susceptible to becoming novel therapeutic targets in diffuse midline glioma. Diffuse midline gliomas (DMG) are treatment-resistant and uniformly fatal pediatric brain tumors. The prognosis of this brainstem tumor is dismal with a median overall survival of 9–12 months from diagnosis. In this study, the authors first developed a lab model of the disease and reanalyzed data from the EGA to identify genes potentially linked to tumor severity. This approach revealed involvement of specific metabolic pathways, suggesting new possibilities for therapeutic intervention. In mice, the study shows that the use of statins can improve survival. For a brief overview of the underlying science, diffuse midline gliomas present intratumor heterogeneity with subpopulations of less-differentiated oligodendrocyte precursors and more differentiated astrocytes. Authors established in vitro models to recapitulate both phenotypes and identify metabolic programs in both subpopulations. To determine the clinical relevance, the authors re-used gene expression data from 76 DMG patients identifying a gene signature predicting decreased overall survival. After extensive metabolic characterization of subpopulations, authors defined strategies to target specific metabolic vulnerabilities. Pre-clinical experiments in mice using OXPHOS inhibitors and statins showed a reduction in tumor burden and increased overall survival [4]. In this scenario, the re-use of high-quality patient’s sequencing data appeared as an advantageous strategy to support the clinical relevance of in-vitro and pre-clinical findings. On the other hand, in rare conditions such as childhood cancer, the availability of public datasets and the possibility of pooling data from different sources can be the only way to obtain impactful results that lead to improvements for patients. References Davidoff, A. M. Pediatric Oncology. Seminars in pediatric surgery 19, 225–233 (2010). Gröbner, S. N. et al. The landscape of genomic alterations across childhood cancers. Nature 555, 321–327 (2018). Venu Thatikonda et al. Comprehensive analysis of mutational signatures reveals distinct patterns and molecular processes across 27 pediatric cancers. Nature Cancer 4, 276–289 (2023). Mbah, N. E. et al. Therapeutic targeting of differentiation-state dependent metabolic vulnerabilities in diffuse midline glioma. Nature Communications 15, (2024). Datasets and Studies ID Title Access Policy Year EGAD00001000134 Driver mutations in histone H3.3 and chromatin remodelling genes in paediatric glioblastoma McGill-DKFZ Pediatric Brain Tumour Consortium 2016 EGAS0000100192 Somatic Histone H3 Mutations in Diffuse Intrinsic Pontine Gliomas and Non-Brainstem Paediatric Glioblastomas St. Jude Children's Research Hospital - Washington University Pediatric Cancer Genome Project 2017 EGAS00001000575 Whole genome sequencing and whole exome sequencing of DIPG tumors and matched normal tissue The hospital for sick children ("SickKids") 2016 EGAD00001000792 Exome sequencing reads of paediatric glioblastoma McGill-DKFZ Pediatric Brain Tumour Consortium 2016 EGAD00001002006 Whole genome sequencing of paediatric glioblastoma in the ICGC PedBrain project ICGC PedBrain project 2016

Blog drug-development-for-pediatric-cancers

Highlights from 2025: Our top 10 requested studies

In 2025, as in previous years, EGA has facilitated the reuse of numerous datasets spanning a broad range of research and healthcare topics. To reflect on this activity, we would like to highlight the top 10 studies behind the most requested datasets of last year. Notably, the vast majority of these highly requested datasets are cancer-focused, particularly involving the study of immune checkpoint inhibitor therapies. Immune checkpoint inhibitors (ICIs) have become one of the most transformative advances in modern cancer therapy. These monoclonal antibodies enhance antitumor immune responses by targeting inhibitory immune checkpoint molecules. While ICIs have led to remarkably favorable and durable responses in some patients, treatment outcomes vary widely. This variability has fueled intense research efforts aimed at understanding resistance mechanisms, exploring combination therapies, and discovering predictive biomarkers to better stratify patients and guide treatment decisions. Study Title Disease 1 EGAS00001005503 Integrated genomic analyses reveal molecular correlates of clinical response and resistance to atezolizumab in combination with bevacizumab in advanced hepatocellular carcinoma Liver Cancer 2 EGAS00001005013 Intratumoral plasma cells predict outcomes to PD-L1 blockade in non-small cell lung cancer Lung Cancer 3 EGAS00001004809 BIOKEY: A single-cell catalogue of the dynamic changes underlying Checkpoint Immunotherapy response in Early Breast Cancer Breast Cancer 4 EGAS50000000138 BIOKEY: Immune heterogeneity in small cell lung cancer and vulnerability to immune checkpoint blockade Lung Cancer 5 EGAS50000000497 Bladder cancer subtyping study across 4 atezo clinical trials Bladder Cancer 6 EGAS50000000689 MOSAIC - Multi-Omics Spatial Atlas In Cancer Cancer Atlas 7 EGAS00000000083 METABRIC (Breast cancer - genome and transcriptome) Breast Cancer 8 EGAS00001005702 Human genomic and phenotypic synthetic data for the study of rare diseases Synthetic data 9 EGAS00001002556 TGF-β attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells Bladder Cancer 10 EGAS00001004353 Molecular subsets in renal cancer determine outcome to checkpoint and angiogenesis blockade Renal Cancer The most requested dataset of the year was EGAD00001008128. It belongs to a study that analyzed 358 liver cancer patients enrolled in clinical trials and treated with the combination of atezolizumab (anti–PD-L1) and bevacizumab (anti-VEGF), atezolizumab alone, or sorafenib. The study highlights how bevacizumab enhances anti-PD-L1 effect and identifies candidate biomarkers to predict treatment response. Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer and the third leading cause of cancer-related death worldwide [1]. Because symptoms often appear late, many patients are diagnosed at advanced stages, when curative options are no longer viable. For many years, sorafenib—a multikinase inhibitor approved in 2007—was the only systemic treatment available. Although numerous clinical trials followed, it was not until a decade later that additional therapies showed meaningful clinical benefit. Among these advances, immunotherapy emerged as a promising first-line option [2]. In particular, the combination of atezolizumab and bevacizumab demonstrated improved survival in patients with advanced HCC, reshaping the treatment landscape. However, treatment efficacy remains limited by the complex and diverse hepatic microenvironments that influence immune responses, as well as by high intratumoral heterogeneity. Identifying robust predictors of treatment response is critical to enable effective patient stratification and maximize the benefits of these therapies [2]. The continued demand for these datasets highlights the interest and critical role of data reuse in deepening our understanding of cancer, accelerating the development of more effective therapies, and improving patient outcomes. By providing secure access to high-quality genomic and clinical data, EGA acts as a catalyst for biomedical progress. References Sung, Hyuna, et al. "Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries."CA: a cancer journal for clinicians 71.3 (2021): 209-249. Ladd, Alexandra D, et al.“Mechanisms of drug resistance in HCC.” Hepatology 79(4) (2024): 926-940

Blog top-10-studies-2025

cBioportal lands at EGA: What this means for researchers

cBioPortal is a widely used platform designed to simplify the exploration of complex cancer genomic data. It brings together large-scale datasets—covering mutations, copy number alterations, gene expression, and clinical annotations—and presents them through an intuitive, interactive interface. Since its launch, it has become a cornerstone of cancer research, helping translate raw genomic data into meaningful biological insights. cBioPortal Is Now Available at EGA At the European Genome-phenome Archive (EGA), we host controlled-access genomic and phenoclinical datasets. By integrating cBioPortal within EGA, researchers could explore cancer datasets (once access has been granted by data owners) through the powerful and user-friendly cBioportal interface. We believe this marks an important step forward for the cancer research community, with several key benefits: Faster Insights Researchers can quickly visualize and explore datasets through the cBioPortal interface before downloading and analyzing complex raw data. This provides an immediate overview, saving time and enabling more informed decisions about which datasets to pursue. Broader Accessibility Working with complex genomic data often requires advanced computational and bioinformatic expertise. This integration lowers the technical barrier, allowing a wider range of scientists to engage with high-value datasets and, for example, accelerate hypothesis generation. Enhanced Data Integration cBioPortal enables the seamless combination of genomic and clinical data, as well as the comparison and exploration of multiple datasets within a single platform. This promotes more comprehensive analyses and supports deeper biological insights. The arrival of cBioPortal at EGA brings together powerful visualization and controlled data access, making it easier than ever for researchers to transform complex cancer genomics data into actionable knowledge. Data submitters: Would you like to make your cBioPortal data available? See our documentation page.

Blog cbioportal

Epigenetic Heterogeneity in Breast Tumors with Implications for Disease Diagnosis and Monitoring

This study investigates epigenetic heterogeneity in human breast cancer using complementary single-cell and bulk epigenomic approaches. Human breast tumor samples were profiled using whole-genome bisulfite sequencing, enzyme-mediated methylation sequencing of sorted tumor-cell populations, single-nucleus CUT&Tag, and single-cell RNA and chromatin-accessibility multiome profiling. These data were used to characterize recurrent epigenetic cell states and intra-tumor heterogeneity in breast cancer.

Study EGAS50000002070

NGS in FACSorted populations from AML patients

FACS-enriched bone marrow cell populations from 22 AML patients were analyzed using a custom panel targeting 56 genes recurrently mutated in myeloid neoplasms. DNA was released from a median of 60,000 sorted cells by incubation with 2.5 μL proteinase K and 2 μL RNase A (Thermo Fisher Scientific) for 30 minutes at 37°C, purified using 2× AMPure XP beads (Beckman Coulter, Brea, CA), and stored in pH 8 elution buffer at −20°C. DNA quantity was measured using the Qubit dsDNA BR Assay on a Qubit 3.0 fluorometer (Thermo Fisher Scientific), and quality was assessed using the Genomic DNA ScreenTape assay on a 4200 TapeStation system (Agilent Technologies, Santa Clara, CA). Libraries were prepared from at least 20 ng of DNA according to the SOPHiA GENETICS protocol and sequenced on an Illumina MiSeq using 2 × 251-cycle paired-end reads. At least 95% of target bases achieved ≥1,000× coverage. FASTQ files were analyzed in SOPHiA DDM using the GRCh37/hg19 reference genome for alignment, variant calling, and annotation. Variants were retained when the variant allele frequency was ≥2%, coverage was ≥100 reads, and the alternate allele was supported on both strands. Synonymous, intronic, intergenic, 5′ untranslated-region, and recurrent panel-artifact calls were excluded. The remaining variants were classified according to Association for Molecular Pathology and American College of Medical Genetics and Genomics guidance using VarSome, ClinVar, COSMIC, and dbSNP. Only pathogenic or likely pathogenic variants were analyzed. All retained calls were visually confirmed in Integrative Genomics Viewer v2.11.3, and only variants previously reported as somatic in hematopoietic or myeloid neoplasms were retained.

Study EGAS50000002132

8432 results for "canc*"

in 19.03 milliseconds.