Expert Panel on Data Challenges in Translational Research
1 like358 views
The document discusses an expert panel on data challenges in translational research within pharmaceutical and biomedical organizations, focusing on the importance of data-driven decision-making to enhance innovation success rates. Key capabilities for a translational medicine platform include data ingestion, patient prioritization, comparative analysis, and visualization of insights. It highlights challenges such as user experience, data interoperability, and the need for effective data governance to leverage healthcare data efficiently.
![Based in Cambridge, UK since 2008, on the Wellcome Genome Campus
AI-augmented knowledge discovery platform for Life Sciences R&D
• Human & animal health
• Personal care and cosmeceuticals
• Food and nutraceuticals
Delivering the innovation platform for the genomics era:
e[automateddatascientist]
to enable data driven decisions
to increase the success rate of innovation
About Eagle Genomics](https://image.slidesharecdn.com/translational-webinar-slides-171215105311/85/Expert-Panel-on-Data-Challenges-in-Translational-Research-3-320.jpg)
![e[automateddatascientist] in translational medicine
Framework applicable in other areas such as preclinical research, animal health, personal care, agri-tech
[to bedside] Interventional Observational
Application Therapeutic
targets
Diagnostic
biomarkers
Biological
mechanisms
Curate patient
data
Select
cohorts
Compare
cohorts
Pinpoint
insight
Patient data
[from bench]
Medical history Lab results
Biomolecular
assays
Images
Internal Collaborative
e[automateddatascientist]
e[curate]
Build
catalog
e[catalog] e[discover] e[discover]+e[nsembl]e[hive]
Capability 1 Capability 4Capability 3Capability 2
Drug
repurposing
…](https://image.slidesharecdn.com/translational-webinar-slides-171215105311/85/Expert-Panel-on-Data-Challenges-in-Translational-Research-6-320.jpg)
![Capability 1:
ingest/curate/enrich
value-based curation and
cataloguing
Capability 2:
prioritise/select
prioritisation selector
Capability 3:
comparison analysis
workflow management
system and analysis
pipeline builder
Capability 4:
visualise/pinpoint
insight navigator
&
genome browser
e[automateddatascientist]
Eagle’s platform - the e[automateddatascientist] has each capability embedded in
a module or a set of modules.](https://image.slidesharecdn.com/translational-webinar-slides-171215105311/85/Expert-Panel-on-Data-Challenges-in-Translational-Research-7-320.jpg)
Expert Panel on Data Challenges in Translational Research
- 1. Expert Panel on Data Challenges in Translational Research in Pharmaceutical and Biomedical Organisations 13.12.2017 – 3PM GMT
- 2. Xose Fernandez Chief Data Officer Panelists Alexandre Passioukov VP Translational Medicine Abel Ureta-Vidal Founder & CEO
- 3. Based in Cambridge, UK since 2008, on the Wellcome Genome Campus AI-augmented knowledge discovery platform for Life Sciences R&D • Human & animal health • Personal care and cosmeceuticals • Food and nutraceuticals Delivering the innovation platform for the genomics era: e[automateddatascientist] to enable data driven decisions to increase the success rate of innovation About Eagle Genomics
- 4. Today: Significant inertia in biologists doing data-rich tasks Data sources meta data summary documents protocols experiment al & clinical reports Concept entailment pathways patients genes supporting evidence drugs Jennifer Biologist
- 5. Capability 1: ingest/curate/enrich ability to ingest/curate/enrich patient in a semi-automatic manner and build a catalog of patients Capability 2: prioritise/select ability to prioritise/select the most valuable/relevant patient or groups of patients Capability 3: comparison analysis ability to run complex comparison analysis between groups of patients, using data Capability 4: visualise/pinpoint ability to pinpoint the new insight generated by the analysis and assess its validity/ robustness/value A translational medicine platform needs to have 4 main capabilities: In order to deliver its promise for new insight discovery:
- 6. e[automateddatascientist] in translational medicine Framework applicable in other areas such as preclinical research, animal health, personal care, agri-tech [to bedside] Interventional Observational Application Therapeutic targets Diagnostic biomarkers Biological mechanisms Curate patient data Select cohorts Compare cohorts Pinpoint insight Patient data [from bench] Medical history Lab results Biomolecular assays Images Internal Collaborative e[automateddatascientist] e[curate] Build catalog e[catalog] e[discover] e[discover]+e[nsembl]e[hive] Capability 1 Capability 4Capability 3Capability 2 Drug repurposing …
- 7. Capability 1: ingest/curate/enrich value-based curation and cataloguing Capability 2: prioritise/select prioritisation selector Capability 3: comparison analysis workflow management system and analysis pipeline builder Capability 4: visualise/pinpoint insight navigator & genome browser e[automateddatascientist] Eagle’s platform - the e[automateddatascientist] has each capability embedded in a module or a set of modules.
- 8. Main hurdles and challenges • User experience for the specific business cases to drive adoption • Interoperability and integration with existing systems • Use and/or creation of standards ie FAIR principles (not a standard) for smooth data/insight exchange • Beyond finding relevant data, representing insight for systematic capture • Federated data governance, involving cultural change
- 9. Data Challenges in Translational Research in Pharmaceutical and Biomedical Organisations Dr Xosé M Fernández Chief Data Officer, Institut Curie Institut Curie
- 11. 11 Current Status Institut Curie Service Transform Health and social care services are under extreme financial pressure. Data is key and underpin the capabilities that will enable insight-based service transformation. Digital Maturity Healthcare information systems generate vast amounts of data that must be used for enabling high-value care. Information ecosystem Integration of data is an increasingly critical goal. Connecting the “system” and enabling data to flow across the organisation. Strategic alignment Our current health and social care service model is changing (e.g. new models of care) moving away from traditional mapping service type and organisation. Analysis capability High-value and effective service delivery and improved outcomes cannot happen without converting rich, high quality, contextual data into insight.
- 12. 12 Tackling some Data Challenges Data creation, access, use and reuse Promote sharing policies enshrined in institutional Data Management Plan Metadata as a means of providing context for datasets, in order to facilitate future discovery, access, aggregation, use and reuse of data Support data creators to submit their data or other research materials to a trusted and sustainable repository, for further curation and long-term preservation, in line with documented collecting policies Identification of digital objects, to facilitate discovery and linking of datasets Provide quality support and resources as a means of building capacity and data skills within the Institut Curie. Institut Curie
- 13. 13 Heterogeneous Data Imaging Radiographic imaging 2D/3D ultrasound Electron microscopy 4D microscopy Histopathology Single-cell imaging… 60% EHRs Dossier médicale personnel Dossier Communiquant en Cancérologie 5% Genomic data Genomic platforms France Médecine Génomique 2025 20% Research data Clinical trials Phenotypes LIMS Multiple research databases… 15% Institut Curie
- 14. 14 Information at your Fingertips Institut Curie *Continuum Soins Recherche *
- 15. 15 Digital Ecosystem Connecting different datasets requires work on descriptors (critical time consuming step), as historical data collections tend to be poorly annotated. METADATA INFRASTRUCTURE A central data hub hosting pseudo-anonymised datasets in line with GDPR† requirements, with links to raw data, as well as a ConSoRe instance for EMR querying. Suitable pilot project is the key for success, it must combine diverse datasets (genomics, transcriptomics, proteins, EHRs) with a team keen to invest time to improve metadata. PILOT PROJECT Data Aggregation – How it works? Institut Curie SEMANTIC LAYER Discoverability is guaranteed by APIs* which aggregate pseudo-anonymised datasets without revealing IDs hosted in secured space. *Application programming interface † General Data Protection Regulation
- 16. Institut Curie Most sequencing will be happening at healthcare and not research By 2025 over 60 million genomes Data geographically distributed Clinical data not interoperable Healthcare not used to handle Terabytes or Exabytes Technical knowhow currently in the research community Secure access and governance Human Genomics in Healthcare
- 17. Institut Curie Main Challenges faced as Chief Data Officer GDPR* as the main data challenge Establish new governance (Curie Data Charter) Outlining policies and procedures (Data Management Plan) Data aggregation to consolidate digital ecosystem Interoperability Cultural change – Information as an asset Translate data into insightful information *General Data Protection Regulation
- 18. Eagle Genomics - Translational Webinar, Dec 13 Introduction – Dr A. Passioukov
- 20. Data Warehouse Data Analytics Knowledge Base Structured data Analysis Mechanistic Insight Genome Etc.. Transcriptome Proteome Metabolome Phenome Reseach & Dev. global network Global Precision medicine-based healthcare system Bioinformatics platform foundational for future healthcare
- 21. SEMI-VIRTUAL TM CONCEPT: COMMON DATABASE TO 3 INNOVATION UNITS (ONCO/DERMA/CNS)
- 22. 1. Use of cutting edge technologies (incl. digital); 2. Semi-virtual structure, with a common DataBase ‘agnostic’ of therapeutic area; 3. Probability-of-Success (PoS) mindedness - focus on priority R&D programs: CDx, go-no-go decisions; 4. Strong synergy with research: using patient data to generate innovation & early inform IUs of winning options; 5. Exploration interfaces between IUs - identification of niches based on molecular biology Pierre Fabre TM: key principles
- 23. Stronger Together & with Others Turn to External OpportuniCes/CollaboraCon In-Silico Pla5orm to connect to external collabora.ons Acquire Data and/or Generate Data Data
- 24. BIG DIGITAL CHALLENGE FOR BIOPHARMA 1. Lack of vision for business value creation potential 2. Data access – data silos 3. Formats / Standards / Dictionaries in a heterogeneous “data lake” 4. Weak signals detection for competitive advantage