Scientific Culture and Accountability Plan (SCAP)

Data Provenance, Data Integrity, Scientific Rigor, and Science Culture Action Plan
Department of Pharmacology and Cancer Biology

Overview: Research in the Department of Pharmacology and Cancer Biology addresses a myriad of important scientific questions relevant to cancer, metabolism, nervous system function, drugs of abuse, and environmental toxicants. One unifying theme that has emerged in the department is cell signaling and the use of pharmacological approaches to define the key regulatory steps in relevant pathways that may be amenable to pharmaceutical exploitation or whose dysregulation may be involved in the pathogenesis of disease.

The challenge: The research enterprise is a continuum that starts with the genesis of an idea, design and implementation of approaches to collect/generate data, analysis and interpretation of processed data and reporting of conclusions. At each step there is the potential for errors in scientific design, execution, analysis and reporting all of which can impact the integrity of the research. Each step is also vulnerable to the biases of the investigator and his/her colleagues.  Pressures to publish in high impact journals, obtain funding to be competitive in the job market, obtain research grants and awards and the actions of competitors, can also impact scientific judgment. Mitigation of the impact of such influences on the rigor and reproducibility of research is the responsibility of everyone in the department. Science is generally self-correcting in that new insights/advances are usually reinforced by the findings of others. However, research that is not performed with the highest level of rigor results in wasteful efforts to reproduce findings and such efforts negatively impact progress.

Responsibility: In our department each laboratory sets its own research agenda and goals and is responsible for ensuring the integrity and provenance of research data. However, it is the responsibility of each member in the department (faculty, staff, fellows and students) to make sure that they actively participate in the establishment and maintenance of a work environment that fosters high quality rigorous research. Every person in the department should feel comfortable expressing their opinions on research findings and their suggestions should be entertained in a respectful, collegial manner. It is the responsibility of the principal investigator/laboratory head to ensure that best practices to promote these guiding principles are implemented and adhered to.

Overview of general principles and best practices (discipline independent)

  • Scientific integrity and the provenance of data is the collective responsibility of all faculty, staff, fellows and students.
  • Projects/experiments should be undertaken with the goal of rigorously testing a hypothesis and not designed to “prove a hypothesis”. Descriptions of research in presentations and publications should be balanced and accommodate the possibility of alternative interpretations and conclusions.
  • It is expected that in the development of new research projects, or the design of experiments, scientists should be aware of and consider the relevant current literature. Knowledge of what has been done in the past by others in the laboratory in the specific area of interest is also a fundamental responsibility of a researcher.
  • All details of an experiment should be recorded in such a manner that it can be reproduced by others and that the primary, unmanipulated, data are accessible for re-analysis. Such recordings should include a description of the rationale for the experiment, the experimental steps used, how data were collected and where they can be located, how the data were analyzed and a discussion of how the data were interpreted.
  • Researchers have a responsibility to ensure that data from the laboratories of collaborators are performed in a rigorous manner and appropriately recorded. All primary data must be accessible to all members of a collaborative team.
  • All researchers should be aware of what constitutes a conflict of interest (financial, intellectual and personal) and should ensure that such conflicts are managed and recorded appropriately.
  • Scientists should be encouraged to present at professional meetings and events and all should be receptive towards, and feel empowered to participate in, constructive discussion of research findings and methodologies.
  • Mechanisms should be established to allow concerns about integrity of data and/or the manner in which studies are performed to be discussed.

Consideration of Best Practices as they relate to our discipline

Experimental design and execution

  • Unbiased hypothesis testing is a central principle that must be considered in the design of all experiments.
  • Investigators have a responsibility to be trained and/or have established competency with techniques/equipment that will be used.
  • All experiments should include the relevant positive and negative controls needed to validate/explain and allow interpretation of new findings.
  • Validate reagents, or be aware of the validation strategy used, by vendors and collaborators. This could include sequencing of plasmids, testing of an antibody with appropriate controls or confirmation of the identity of cell lines and the purity of a chemical/drug.
  • A statistical plan, where appropriate, and predefined endpoints should be established and recorded before experimentation begins. If deviations from protocol are required they must be recorded along with an explanation of the rationale why such deviations were deemed necessary.
  • Where possible (and practical) blinding should be used to insulate the investigator to potential biases in measurement. This is particularly applicable to experiments where exact measurements are not possible (i.e. tumor measurements, gating in flow cytometry, staining intensity in immunohistochemistry) or where the experimental observations are quantified by an observer rather than an instrument or computer.

Analysis of experimental results

  • Experimental design issues should be resolved before experiments are conducted: power analysis to determine appropriate sample size, determination of independent and dependent measures and confounds, set and defend technical criteria for data inclusion/exclusion.
  • Conclusions should be tested by rigorous statistical procedures.
  • Statistical tests should be selected in advance of the study as appropriate to the design and biological question being asked.
  • Even if a biostatistician is consulted for data analyses, the investigator is ultimately responsible for understanding the meaning and interpretation of the statistical tests.

Storage and provenance of experimental data

  • All primary data should be stored and appropriately indexed to allow access in the future by the investigator or others. Any and all transformations of primary data should be recorded and date stamped with an explanation of the reason for the modification(s).
  • Published data, including both primary and processed data, should be archived in a read-only directory on an enterprise class storage system (e.g. Duke Isilon, Google Cloud Storage, Amazon S3, etc.) with a checksum calculated to ensure data integrity.
  • Primary genomic, metabolomic, and proteomic data should be submitted and archived with field specific data repositories (e.g. NCBI Geo, Mendeley Data, etc.).
  • Processing of large datasets should be scripted when possible to ensure reproducibility. All software tools and scripts developed in the course of a project should be under version control (e.g. GitHub).

Implementation of this plan

  • Every individual in the department engaged in research is responsible for familiarizing themselves with this plan and should conduct their research in a manner that is compatible with its contents.
  • Each principal investigator/lab head/core director should discuss this plan with every member of their research group and provide verification that full compliance has been achieved. The plan should be discussed with new hires before they commence research in the laboratory.
  • All researchers are responsible for keeping their conflict of interest form up to date. Details on how to complete this form and information on who are required to fill such disclosures are available at https://medschool.duke.edu/research/ethics-integrity-compliance/research-integrity-office. As a general rule anyone who could benefit financially from research ongoing in a laboratory should complete this form.
  • An essential component of compliance with this policy is verification that every researcher has completed the new responsible conduct in research requirements that have been implemented. Details as to how this requirement can be fulfilled can be found at https://medschool.duke.edu/research/ethics-integrity-compliance/responsible-conduct-research-program.

Other useful resources and contacts

  • Duke Health Integrity Line: (800) 826-8109
  • Duke University Compliance and Fraud Hotline: (800) 849-9793
  • HIPAA Privacy Line: (800) 688-1867
  • Animal Welfare Hotline: (919) 684-3535

Contacts: Any questions, comments or suggestions that could help to improve this plan or its implementation should be directed to Donald McDonnell, PhD, (Donald.Mcdonnell@duke.edu). This plan has been developed in consultation with the Advancing Scientific Integrity, Services and Training (ASIST) office. Further resources and information on this initiative can be obtained at https://medschool.duke.edu/research/research-support-offices/advancing-scientific-integrity-services-and-training-office or by contacting the office directly at asistoffice@duke.edu.