NIH Genomic Data Sharing (GDS) Policy

Kathleen Calzone, PhD, RN, APNG, FAANAnjan Purkayastha, PhD

December 2, 2015

• Brief Overview• Effective Dates and Timelines• Scope & Applicability• Data Standards• Repositories• The Data Sharing Process• Exceptions to GDS policy• Governance and Resources for GDS• Q&A

Outline

Brief Overview

GDS Policy

• Broad data sharing promotes maximum public benefit from federally funded genomics research

• NCI supports the broadest appropriate genomic data sharing with timely data release through broadly accessible open or, if more appropriate, controlled access data repositories

• Systems to ensure robust participant protection and appropriate oversight of research conduct, data quality, data management, data sharing, and data use are fundamental to effective data sharing policies and practices

• Data sharing allows data generated from one research study to be used to explore a wide range of additional research questions. It also enables data from multiple projects to be combined, amplifying the scientific value of data

• Data to be shared should be annotated to enable data reuse, understanding, harmonization and meta-analysis

Overarching Principles

• The GDS Policy takes effect for competing grant applications and contract proposals submitted to NIH on or after January 25, 2015.

• The first round of grant applications submitted for the January 2015 cycle would likely receive funding in the late Summer or early Fall of 2015.

• To align with this the intramural programs were directed to comply with GDS on data generated on or after August 31, 2015.

Genomic Data Sharing Policy Effective Dates

• Within CCR, GDS will be implemented in phases.

• In the first phase all Human studies will implement the GDS policy effective December 1, 2015.

• For ongoing Human studies Genomic Data Sharing Plans should be submitted by December 31, 2015.

• Implementation of GDS policy across non-human studies will start in the second phase, date TBD.

Genomic Data Sharing Policy Effective Dates

Scope and Applicability

GDS Policy

The GDS Policy applies to all NIH-funded research that generates large-scale human or non-human genomic data as well as the use of these data for subsequent research

Large-scale data include data from genome-wide association studies (GWAS) and single nucleotide polymorphism (SNP) arrays, as well as genome sequence, transcriptomic, metagenomic, epigenomic, and gene expression data, irrespective of funding level and funding mechanism

GDS Scope

Thresholds for Data Sharing Expectations

* Includes distinct individuals, species, strains, samples, treatments, time points, tissues, etc. Example: Data from 25 patients at 4 time points after treatment would reach a 100 sample threshold; Data from 50 tumor-normal comparisons would reach a 100 sample threshold.

Shareable Data: Additional Criteria

Examples of smaller-scale projects that the NCI anticipates data sharing for (regardless of study design) include, but are not limited to: Projects examining rare cancers or rare cancer related outcomes, rare

cancer subtypes.

Projects focusing on under-studied populations.

Based on NIH IC priorities, GDS policy may apply to smaller data-sets: State of science NCI’s programmatic priorities Utility of data to research community

Journal(s) may require data sharing before publishing your manuscript

Examples of Research Outside the Scope of the GDS Policy

Examples of projects outside the Policy’s scope include, but are not limited to, those that do not meet the criteria in the above examples and involve:X Instrument calibration exercises.X Statistical or technical methods development.

Genomic Data Sharing versus Human Data Sharing Policies: Similarities and Differences

GDS HDS

Implementation Date

8/31/2015 10/1/2015

Scope Large scale genomic data in the NIH Extramural and Intramural Research Programs

All human data in the NIH Intramural Research Program at the time of pre-research scientific review

Data Sharing Plan Required Required

Exceptions Infrastructure for approval in place

Not established

Databases for Submission

Established Not specified

Protocol and Consent

Protocol and consent language approved by IRB.

Protocol language approved by IRB.

Data Standards

GDS Policy

Minimal Information Guidelines

• Data metadata around the experiment or study and annotations that are necessary to reproduce any published table or analysis must be included with genomic data submissions. This includes all relevant study design, materials and methods, and analytic methods.

• The specimen acquisition and experimental procedures as well as the data processing and analysis methods (such as alignment algorithms, software versions, etc.) are required with data submission.

• Data pertinent to the interpretation of genomic data—such as associated phenotype data (e.g., clinical information), exposure data, relevant metadata, and descriptive information (e.g., protocols or methodologies used)—are expected to be shared.

Human and Non-Human Genomic Data Repositories

dbGaP

Cancer Genome Hub

Human studies Non-human studies

Gene Expression Omnibus (GEO)

WormBase

Zebrafish Model (ZFIN)

Mouse Genomic Informatics (MGI)

Array Express

Trace Archive

Sequence Read Archive

Model Organism Databases

GenBank

European Nucleotide Archive (ENA)

DNA Data Bank of Japan (DDBJ)

http://gds.nih.gov/02dr2.html

19

The Data Sharing Process

GDS Policy

Key Documents: Genomic Data Sharing Plan (GDSP)

The NCI expects that GDSPs will be collected and reviewed at earliest time-point possible. Components of a GDSP• Each project/protocol requires a separate GDSP• Elements of a Genomic Data Sharing Plan consist of the following:

Human and/or non-human Project/protocol Information including sample size Sample and data types Data repository identified for submission Data submission and release timeline Data use limitations

• For studies involving human data, the additional elements below will be considered in the development of data sharing plans:

Research participant informed consent [including (IRB)] assessments of informed consent processes and consent documents with regard to broad data sharing and future research use

Participant protection concerns for the particular study population (e.g., participant privacy or potential for group harm) or scientific design (e.g., isolated geographic population or small family studies).

Genomic Data Sharing Plan contd. CCR SOP-RPS-21 Establishing a Genomic Data Sharing Plan• Details GDSP procedures• Fillable GDSP form with electronic signature

The SD, or their delegate, and GPA should review DSP at the earliest time point possible. • Clinical Trial -The DSP should be submitted to, and reviewed

by, the SD (or delegate) and GPA at the time of IRB review and as approval is being made. PSO working on study amendments for open protocols.

• Other Research – The GDSP should be submitted to, and reviewed by the SD (or delegate) and GPA before funding decisions are made (Prospective Review) or prior to data generation (Retrospective Review). Ongoing studies must submit a GDSP if data generated on or

after August 31, 2015.

Key Documents: Institutional Certification

• Institutional Certification (IC), assures plans for the submission of human genomic data meet the expectations of the Genomic Data Sharing Policy.

• An IC must accompany the submission of all human data to the NIH Database of Genotypes and Phenotypes (dbGaP) at the time of study registration.

• Submit the IC to the Genomic Program Administrator (GPA) in accord with all terms outlined in the CCR SOP-RPS-23 Study Registration to dbGAP. – Explicit informed consent for open access data sharing should be

provided by research participants for studies depositing full genome sequence into open access, whether sample or data collection is prospective or retrospective;

Institutional Certification contd.– For studies that initiate participant recruitment after the

implementation date of the NIH GDS Policy, should assure that future research use and data sharing are consistent with the informed consent provided by study participants.

– For studies where participant recruitment occurred prior to the effective date of the NIH GDS Policy, should assure that future research use and data sharing are not inconsistent with the informed consent provided by study participants.

– If established or commercially available cell lines or clinical specimens are included as data sources within the study, investigators should seek whenever possible to use sources where consent for future research use and data sharing can be documented.

– In all cases, and consistent with program priorities, phenotype or clinical variables submitted for data release may be adjusted to promote participant privacy or other participant protection concerns as assessed by an IRB.

Institutional Certification contd.

• Clinical Trials– IC memo(s) should be filled out and signed by the study

PI and Scientific Director (SD) immediately following IRB approval. • This procedure assures that the Institutional Review Board

(IRB) reviewed the informed consent and protocol to assure that that data sharing is appropriate and identified any data use limitations that may exist based on the language found in the consent.

Institutional Certification contd.

• Other Research– IC memo(s) should be filled out and signed by the study PI

and Scientific Director (SD) prior to data generation.

– If applicable, the GPA and/or Institutional Review Board (IRB) will review the informed consent and protocol to assure that that data sharing is appropriate and to interpret any data use limitations that may exist based on the language found in the consent.

• GPA or GPA administrator will register all studies with human genomic data that fall within the scope of the GDS Policy in dbGaP regardless of which NIH-designated data repository will receive the data. - If an exception has been granted by the NCI, the study still

must be registered in dbGaP in accord with this timeline, but data deposition will not be expected.

• For All Shareable Data Sets − Study registration should occur once data cleaning and

QA/QC are complete.− Registration starts with the PI mailing the GPA with a

request to share data; sharing the IC memo as an attachment.

Study Registration

Exceptions to GDS Policy

Requests for Exceptions to Data Sharing

• In rare cases, NCI will consider requests for an exception to usual data submission expectations.

• Submission of genomic data may be precluded by: – international laws; – limitations in the original informed consents; – concerns about harms to individuals or groups; – or other cases where expectations for data submission cannot be met.

• CCR SOP-RPS-22 Requesting a Genomic Data Sharing Exception– Details Exception procedures– Fillable Exception request application with electronic signatures

Requests for Exceptions to Data Sharing, contd.In cases where data submission to a NIH-designated data repository is not appropriate:• Investigators should provide a justification for any data submission

exceptions prior to data generation. – This must also include an alternative plan to share data through other mechanisms

whenever possible.

• The Exception Request form should be signed by – Branch Chief– GPA and/or IRB, if using a consent justification for exception– CCR Scientific Director

• Exception requests (including Statements of Programmatic Priority) will be reviewed by the Trans-NCI Data Sharing Working Group and a recommendation will be made.

• NCI Scientific Program Leaders (SPL) will review the exception request package. When a consensus has been reached, the request will be sent to the NCI Director for signature.

• The request must them be signed by the NIH Deputy Director for Intramural Research after the NCI Director.

Requests for Exceptions to Data Sharing

NCI uses the following criteria to assess exception requests:• Impact of data sharing compliance on scientific merit• Unique resource • High value resource• Regulatory considerations

– Missing consent from patients

• Ethical considerations• NIH data sharing exception precedents• Is there an acceptable alternative data-sharing plan (ADSP)?

– Impact of ADSP on data re-use– Impact of ADSP on data discoverability– Burden– Feasibility

• In all cases where an ADSP is determined to be appropriate, information on how to request access to data and a basic summary of the study and study data will be listed in dbGaP (or other appropriate data repository).

Governance, Infrastructure, Resources

GDS Policy

Governance• Trans-NCI Genomic Data Sharing Working Group

– Provide on-going stewardship and leadership for Institute data sharing policies and their implementation. • Has developed the NCI GDS Implementation Framework

document– Consider exception requests to any aspect of NIH or NCI

data sharing expectations– Address any on-going policy or implementation development

needs, and adjudicating or interpreting any aspect of the NIH/NCI policies and practices.

• NCI Data Access Committees (DACs) – Charged with providing oversight and monitoring of data

access activities and participant protection needs related to all NCI supported data sets.

Genomic Program Administrator (GPA)

• Function as a central point of coordination and information about NCI data sharing activities and implementation of NIH and NCI polices.

• Point of contact regarding the implementation of data sharing expectations in the CCR.

• Work in concert with the NCI Genomic Data Sharing Governance Committee on questions regarding implementation of Institute and NIH policies.

• The GPA will serve as a liaison to the NIH Genomic Data Sharing governance structure through the Technical Standards and Data Submission Steering (TSDS) Committee.

• The CCR GPA is Kathy Calzone, calzonek@mail.nih.gov

GPA Administrator

• Point of contact regarding the logistics of study registration and data submission for CCR.

• Work in concert with the CCR GPA on questions regarding implementation of Institute and NIH policies

• The GPA Administrator will also work in concert with the NCI Genomic Data Sharing Governance Committee on questions regarding implementation of Institute and NIH policies.

• The CCR GPA Administrator is Anjan Purkayastha, anjan.purkayastha@nih.gov

CCR and NCI Infrastructure

• Information on NIH-wide implementation of GDS Policy: https://gds.nih.gov/

• Information on NCI-wide implementation of GDS Policy : http://www.cancer.gov/grants-training/grants-management/nci-policies/genomic-data#policy

• Information on CCR-wide implementation of CCR GDS Policy:http://bit.ly/CCR_GDS

• CCR SOPs have been established: – Developing a GDSP (RPS-21)– Study Registration to dbGAP (RPS-23)– Requesting a GDS Exception (RPS-22)

• Clinical Trial Infrastructure– Protocol template language approved– Modification to iRIS application in place– Template consent language

• Approved by IRB 4/2015

• OHSRP Exemption Infrastructure– Language integrated into application

CCR and NCI Infrastructures, continued

CCR and NCI Infrastructures, continued• Compliance

– Manuscript clearance– Protocol tracking– OHSRP tracking– Z01 annual reporting

• Weakest area is non-human genomic data sharing– Genomic Data Sharing Compliance Committee

• Last meeting 10/6/2015• Purpose is to establish requirements and strategy to support

intramural and extramural compliance with the Genomic Data Sharing policy

• Training Options– Training slide set developed, available for Branch and/or team training

as requested– Working with NCBI to establish training sessions for dbGaP

registration/submission

Questions and Answers

Genomic Program Administrator (GPA)Kathleen Calzone, PhD, RN, APNG, FAAN301-435-0538, calzonek@mail.nih.gov

GPA AdministratorAnjan Purkayastha, PhD301-594-1395, anjan.purkayastha@nih.gov