NIH Data Science Special Interest Group
- 1. Biobanking: a user’s perspective and an overview Jonathan Pevsner, Ph.D. Professor, Dept. of Neurology Kennedy Krieger Institute and Johns Hopkins Medicine Chief Scientific Officer, Sturge-Weber Foundation pevsner@kennedykrieger.org Data Science Forum: NIH Data Science SIG Global Perspective on Biobanking and Access to Samples June 23, 2017
- 2. Conflicts of interest I have no conflicts of interest.
- 3. Outline From genotype to phenotype: a framework Three biobanking examples Postmortem brains from the NIH NeuroBioBank Establishing a rare disease biobank From a large genomics dataset to biobank samples Issues, lessons and principles 1. Usefulness 2. Existing biobanks 3. GUIDs: the importance of labels 4. Data science is integral to biobanking 5. Standards 6. Informed consent 7. Needs and opportunities
- 4. The relationship between genotype and phenotype represents one of the most fundamental and challenging problems in biomedical science. Fundamental framework: genotype to phenotype Genotype Phenotype
- 5. Fundamental framework: genotype to phenotype Genotype Phenotype DNA individual populationorgancellprotein We can provide a framework for this problem.
- 6. Fundamental framework: genotype to phenotype Genotype Phenotype DNA individual populationorgancellprotein We can provide a framework for this problem. RNA pathways circuits
- 7. Fundamental framework: genotype to phenotype Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 …
- 8. Fundamental framework: genotype to phenotype Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 ABCD1 severe childhood disease (ALD) mild adult onset disease (AMN) apparently normal One gene mutation can have different phenotypic consequences: the same ABCD1 mutation may result in severe childhood-onset adrenoleukodystrophy (ALD), milder adult- onset adrenomyeloneuropathy (AMN), or no symptoms.
- 9. Fundamental framework: genotype to phenotype Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 GNAQ melanocytes: uveal melanoma endothelial cells: Sturge-Weber blood: apparently normal One gene mutation can have different consequences: when and where mutations occur is crucial.
- 10. Fundamental framework: genotype to phenotype Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 … one disease phenotype, multiple genetic contributors For almost all diseases (including common diseases such as autism or bipolar disorder) we search for multiple genetic variants that confer risk for a phenotype
- 11. Outline From genotype to phenotype: a framework Three biobanking examples Postmortem brains from the NIH NeuroBioBank Establishing a rare disease biobank From a large genomics dataset to biobank samples Issues, lessons and principles 1. Usefulness 2. Existing biobanks 3. GUIDs: the importance of labels 4. Data science is integral to biobanking 5. Standards 6. Informed consent 7. Needs and opportunities
- 12. A port-wine birthmark affects about 1:300 people. It varies in size and location. Sturge-Weber syndrome affects < 1:20,000 people. It affects a subset of individuals with a facial PW birthmark. A user’s perspective on biobanking: three examples. (1) Sturge-Weber syndrome and a brain bank
- 13. A user’s perspective on biobanking: three examples. (1) Sturge-Weber syndrome and a brain bank DNA from blood (presumed unaffected) DNA from port- wine birthmark (presumed affected)
- 14. A user’s perspective on biobanking: three examples. (1) Sturge-Weber syndrome and a brain bank DNA from blood (presumed unaffected) DNA from port- wine birthmark (presumed affected) sequence the genome sequence the genome compare We identified a mosaic mutation in GNAQ as causing Sturge- Weber syndrome and port-wine birthmarks (NEJM, PMID 23656586).We analyzed samples from 3 individual patients.
- 15. A user’s perspective on biobanking: three examples. (1) Sturge-Weber syndrome and a brain bank Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 GNAQ After finding the GNAQ mutation we turned to the NIH NeuroBioBank at the University of Maryland.We obtained 97 samples to validate our findings.The availability of these samples from a biobank was crucial!
- 16. A user’s perspective on biobanking: three examples. (2) Establishing a Sturge-Weber syndrome biobank I am Chief Scientific Officer of the Sturge-Weber Foundation.We need to create (or join) a biobank. • Patients and families tell me “I want to donate my brain and body to science. Can you help?”What’s the plan; and are there informed consent issues? • Scientists have discovered that the GNAQ mutation occurs primarily in endothelial cells, and cell lines have been established from brain biopsies. How can researchers share and access these cell lines? • Are there standards that we should follow in describing the genotype and phenotype of Sturge-Weber syndrome samples and patients? • Have these problems been addressed by those studying related diseases?
- 17. A user’s perspective on biobanking: three examples. (2) Establishing a Sturge-Weber syndrome biobank Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 GNAQ It’s important to link clinical data (e.g. from a patient registry) with data generated from biospecimens!
- 18. A user’s perspective on biobanking: three examples. (2) Establishing a Sturge-Weber syndrome biobank Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 GNAQ What information do we need to capture about cell lines, brain, and skin samples?
- 19. A user’s perspective on biobanking: three examples. (2) Establishing a Sturge-Weber syndrome biobank Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 GNAQ What information do we need to capture about cell lines, brain, and skin samples? How do we relate genomic DNA sequence findings, RNA-seq, proteomics to the samples?
- 20. A user’s perspective on biobanking: three examples. (2) Establishing a Sturge-Weber syndrome biobank Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 GNAQ What information do we need to capture about cell lines, brain, and skin samples? What information do we need to capture about the phenotypes as we collect samples at diverse sites? How do we relate genomic DNA sequence findings, RNA-seq, proteomics to the samples?
- 21. A user’s perspective on biobanking: three examples. (3) Discovering mosaic mutations in 9000 autism samples We asked whether mosaic mutations occur in autism. By applying to NIH we obtained previously generated whole exome sequence data on 9000 individuals via the Simons Simplex Collection (SSC).We discovered mosaic variation is enriched in children with autism spectrum disorder. To validate our findings, we applied to the Simons Foundation and received approval to obtain DNA from a Rutgers repository (http://www.rucdr.org/).We purchased 300 DNA samples and successfully validated our findings. See PMID 27632392:
- 22. A user’s perspective on biobanking: three examples. (3) Discovering mosaic mutations in 9000 autism samples Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 … A user starts with genomics data…
- 23. A user’s perspective on biobanking: three examples. (3) Discovering mosaic mutations in 9000 autism samples Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 … …then purchases cell lines or DNA or brain chunks for further studies… A user starts with genomics data…
- 24. A user’s perspective on biobanking: three examples. (3) Discovering mosaic mutations in 9000 autism samples Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 … …then purchases cell lines or DNA or brain chunks for further studies… Obtaining clinical phenotypes from the biobank is essential. A user starts with genomics data…
- 25. Outline From genotype to phenotype: a framework Three biobanking examples Postmortem brains from the NIH NeuroBioBank Establishing a rare disease biobank From a large genomics dataset to biobank samples Issues, lessons and principles 1. Usefulness 2. Existing biobanks 3. GUIDs: the importance of labels 4. Data science is integral to biobanking 5. Standards 6. Informed consent 7. Needs and opportunities
- 26. (1) Usefulness • Diseases are considered rare when affecting 200,000 or fewer people (U.S. definition) or fewer than 1:2,000 people (European definition). • There are ~6,800 rare diseases. • Biobanks offer crucial resources to help solve the causes of rare diseases—and to study diagnosis, prevention, and treatment. • Biobanks offer a range of cell, solid tissue types (e.g. brain, heart, fibroblasts, lymphoblastoid cell lines) and bodily fluids. • Biobanks offer biospecimens from individuals, pedigrees, and/or populations. • Samples from biobanks are complemented by phenotypic and genotypic data.
- 27. List of panelists Jonathan Pevsner, Professor, at the Dept. of Neurology, Kennedy Krieger Institute. Presentation title: Biobanking user’s perspective and an overview Dept. of Psychiatry and Behavioral Sciences, Johns Hopkins Medicine David van Enckevort, Project Manager BBMRI & RD-Connect,Department of Genetics, University Medical Center Groningen (UMCG). Presentation title: “FAIR (Findable, Accessible, Interoperable and Reusable) data and sample access “ Manuel Posada de la Paz. Director, Research Institute for Rare Diseases (Instituto de Investigación en Enfermedades Raras), a member of the EuroBioBank. Presentation title: Rare diseases biological samples: small collections and research. Kerry Wiles, Program Director- VUMC Tissue Repository, CHTN (Cooperative Human Tissue Network) Western Coordinator. Presentation title: An academic prospective procurement repository: From Donor to Bench Jim Vaught Editor-in-Chief, Biopreservation Journal, past President of the International Society for Biological and Environmental Repositories (ISBER), on the board of directors for ISBER and NDRI (National Disease Research Interchange), Presentation title: "NIH and ISBER perspectives on specimen locators" Daniel Catchpoole Director of Kids Research Institute, The Children's Hospital at Westmead (Australia). Presentation title: The Australian experience, issues and solution
- 28. (2) Examples of existing biobanks and biobank initiatives Coriell Biorepository The NIGMS collection has >11,000 cell lines and ~6,000 DNA samples. https://catalog.coriell.org/ NIH NeuroBioBank 6 sites.The University of Maryland Brain &Tissue Bank has distributed 35,000 tissue samples to >900 researchers. https://neurobiobank.nih.gov/ Cooperative Human Tissue Network (CHTN) Supported by the National Cancer Institute https://www.chtn.org/
- 29. EuroBioBank 130,000 samples available; 13,000 collected and >7,000 samples distributed per year. http://www.eurobiobank.org/ RD-Connect "An integrated platform connecting databases, registries, biobanks and clinical bioinformatics for rare disease research.” http://rd-connect.eu/ Research Institute for Rare Diseases (Instituto de Investigación en Enfermedades Raras), a member of the EuroBioBank. http://www.eurobiobank.org/en/partners/description/ isciii.htm (2) Examples of existing biobanks and biobank initiatives
- 30. BBMRI-ERIC Biobanking and biomolecular resources research infrastructure-European Research Infrastructure Consortium. http://www.bbmri-eric.eu/BBMRI-ERIC/common-service-it/ Kids Research Institute,The Children's Hospital at Westmead (Australia) http://www.kidsresearch.org.au/our-facilities/bio-banks National Disease Research Interchange (NDRI) The mission of NDRI is to provide human biospecimens to advance biomedical/bioscience research and development worldwide.” http://ndriresource.org/ (2) Examples of existing biobanks and biobank initiatives
- 31. All of Us “The All of Us Research Program seeks to extend precision medicine to all diseases by building a national research cohort of one million or more U.S. participants.” It includes a biobank. https://allofus.nih.gov/about/program-components NIMH Repository and Genomics Resource (NIMH-RGR) “…plays a key role in facilitating psychiatric genetic research by providing a collection of over 150,000 well characterized, high quality patient and control samples from a wide-range of mental disorders.” https://www.nimhgenetics.org/ (2) Examples of existing biobanks and biobank initiatives
- 32. (3) GUIDs: the importance of labels “Accession numbers” are alpha-numeric characters that provide links to various kinds of data or records. For example, NP_620258.1 is an accession number corresponding to a protein sequence. A GUID is a Global Unique Identifier that corresponds to a study participant.The GUID facilitates tracking patient’s data across studies and location and over time in a deidentified manner. Example 1: a participant was recruited twice (years apart) to a single study. Example 2: a trio was recruited into two separate autism genome sequencing studies (one study excluded severe phenotypes, one excluded mild phenotypes).The proband’s phenotype had become severe over time.
- 33. (4) Data science is integral to biobanking Biobanking requires a series of tasks such as obtaining biospecimens and associated metadata (e.g. phenotypic data, cause of death, postmortem interval, cell culture conditions, imaging data, genomics data). Goals include effective communication, standardization (e.g. of protocols), and an electronic portal to a repository. All this requires data science.
- 34. Biobanks must integrate diverse data types Genotype Phenotype DNA individual populationorgancellprotein gene1 gene20,000 gene2 … Sequence data: Genomic DNA (dbGaP), RNAseq Proteomics data, metabolomics imaging data phenotypic test data (e.g. neuropsychology) cell culture, biochemical, iPSC data epidemiology
- 35. (5) Standards Biobanks implement Data Dictionaries to manage data elements (and data structures) in a uniform manner.The use of Common Data Elements is crucial. NIH Common Data Elements (CDE) Repository “designed to provide access to structured human and machine-readable definitions of data elements.” https://cde.nlm.nih.gov/home
- 36. Standards: Common Data Elements (CDE) Repository
- 37. (6) Informed consent Research studies (in contrast to clinical tests) are under Institutional Review Board (IRB) jurisdiction.A researcher must have a research protocol approved, and one or more consent forms. Biobanks provide biospecimens that are sometimes in the realm of human subjects research.Appropriate consent forms must be administered for biospecimens to be deposited in a biobank. An emerging issue is obtaining appropriate consent for DNA to be sequenced from biospecimens. Because of the nature of contemporary sequencing samples are no longer inherently deidentifiable.
- 38. (7) Needs and opportunities We need resources and efforts such as the following: • coordination of biobanking efforts across diverse initiatives. • awareness and adoption of community standards for biobanking. • flexibility to adapt to changing technologies (e.g. sequencing technologies). • Integrated platforms and bioinformatics tools