Nothing in this document should be construed to differ from, or substitute for, the policies described in the Federal Regulations for the Protection of Human Subjects [45CFR46 (NIH) and 10CFR745 (DOE)]. Rather, it is intended to supplement those policies by focusing on the particular issues raised by large-scale human DNA sequencing. This statement addresses six topics: (1) benefits and risks of genomic DNA sequencing; (2) privacy and confidentiality; (3) recruitment of DNA donors as sources for library construction; (4) informed consent; (5) IRB approval; and (6) use of existing libraries.
The guidance provided in this statement is intended to afford maximum protection to DNA donors and is based on the belief that protection can best be achieved by a combination of approaches including:
While the HGP offers great promise to humanity, there will be no direct benefit, in either clinical or financial terms, to any of the individuals who choose to donate DNA for large-scale sequencing. Rather, the motivation for donation is likely to be an altruistic willingness to contribute to this historic research effort.
However, individuals who donate DNA to this effort may face certain risks. Information derived from the donors will become available in public databases. Such information may reveal, for example, DNA sequence-based information about disease susceptibility. If the donor becomes aware of such information, it could lead to emotional distress on her/his part. If such health-related information becomes known to others, discrimination against the donor (e.g., in insurance or in employment) could result. Unwanted notoriety is another potential risk to donors. Therefore, those engaged in large-scale sequencing must be sensitive to the unique features of this type of research and ensure that both the protections normally afforded research subjects and the special issues associated with human genomic DNA sequencing are thoroughly addressed.
While some risks to donors can already be identified, the probability of adverse events materializing appears to be low. However, the risks of harm to individuals will increase if confidentiality is not maintained and/or the number of donors is limited to a very few individuals. Either, or both, of these situations would increase the possibility of a donor's identity being revealed without his/her knowledge or permission.
A final issue to consider is characterized in a statement taken from the OPRR Guidebook(1) which points out that "some areas [of genetic research] present issues for which no clear guidance can be given at this point, either because enough is not known about the risks presented by the research, or because no consensus on the appropriate resolution of the problem exists." It is anticipated that the DNA sequence information produced by the Human Genome Project will be used in the future for types of research which cannot now be predicted and the risks of which cannot be assessed or disclosed.
Investigators should introduce as many disconnects between the identity of donors and the publicly available information and materials as possible. There should not be any way for anyone to establish that a specific DNA sequence came from a particular individual, other than resampling an individual's DNA and comparing it to the sequence information in the public database. In particular, no phenotypic or demographic information about donors should be linked to the DNA to be sequenced.(2) For the purposes of the HGP such information will rarely be useful, and recording such information could result in possible misuse and compromise donor confidentiality.
Confidentiality should be "two way." Not only should others be unable to link a DNA sequence to a particular individual, but no individual who donates DNA should be able to confirm directly that a particular DNA sequence was obtained from their DNA sample.(3) This degree of confidentiality will preclude the possibility of re-contacting DNA donors, providing another degree of protection for them. It should be clear to both investigators and to donors that the contact involved in obtaining the initial specimen will be the only contact.(4)
Another approach for protecting all DNA donors is to reduce the incentive for wanting to know the identities of particular donors. If the initial human sequence is a "mosaic" or "patchwork" of sequenced regions derived from a number of different individuals, rather than that of a single individual, there would be considerably less interest in who the specific donors were. Although there may be scientific justification that each clone library used for sequencing should be derived from one person, there is no scientific reason that the entire initial human DNA sequence should be that of a single individual. As approximately 99.9% of the human DNA sequence is common between any two individuals, most of the fundamental biological information contained in the human DNA sequence is common to all people.
To increase the likelihood that the first human DNA sequence will be an amalgam of regions sequenced from different sources, a number of clone libraries must be made available. Although a number of large insert libraries have been made, most do not meet all of the standards set in this document; therefore, these libraries should be used as substrates for large-scale sequencing only under circumscribed conditions (see section 6, below). Starting immediately, new libraries will be developed that have the advantage of being constructed in accordance with the ethical principles discussed in this document; they may also confer some additional scientific benefit. Such libraries are critical for the long-range needs of the HGP.
There are two additional issues that have arisen in considering donor selection. These warrant particular discussion:
Federal regulations (45CFR46 and 10CFR745) require the disclosure of a number of issues in any informed consent document. They include such issues as potential benefits of the research, potential risks to the donor, control and ownership of donated material, long-term retention of donated material for future use, and the procedures that will be followed. In addition, there are several other disclosures that are of special importance for donors of DNA for large-scale sequencing. These include:
Many academic human genetics units have considerable experience in dealing with research subjects and obtaining informed consent, while the laboratories that are likely to be involved in making the libraries for sequencing have, in general, much less experience of this type. Therefore, library makers are encouraged to establish a collaboration with one or more human genetics units, with the latter being responsible for recruiting donors, obtaining informed consent, obtaining the necessary biological samples, and providing a blinded sample to the library maker. Collaboration with tissue banks may be considered as long as these banks are collecting tissues in accordance with this guidance. The library maker should have no contact with the donor and no opportunity to obtain any information about the donor's identity.
It is important that in obtaining consent for contined use of existing libraries, no coercion of the DNA donor occur. It is therefore recommended that consideration be given to whether it is appropriate for the individual who previously recruited the donor to recontact him/her to obtain this consent. In some cases an IRB may determine that the recontact should be made by a third party to assure that the donors are fully informed and allowed to choose freely whether their DNA can continue to be used for this purpose.
Investigators funded by NCHGR and DOE to develop new libraries for large-scale human DNA sequencing will be required to have their plans for the recruitment of DNA donors, including the informed consent documents, reviewed and approved by the funding agency before donors are recruited. Investigators involved in large-scale human sequencing will also be asked to observe those aspects of this guidance that pertain to them.
Francis S. Collins, M.D., Ph.D.
Director, NCHGR, NIH
Aristides N. Patrinos, Ph.D.
Associate Director, OHER, DOE
August 17, 1996
2 It is recognized that it will be trivially easy to determine the sex of the donor of the library, by assaying for the presence or absence of Y chromosome in the library. Return to Text
3 There are a number of approaches to preventing a DNA donor from knowing that his/her DNA was actually sequenced as part of the HGP. For example, each time a clone library is to be made, an appropriately diverse pool of between five and ten volunteers can be chosen in such a way that none of them knows the identity of anyone else in the pool. Samples for DNA preparation and for preparation of a cell line can be collected from all of the volunteers (who have been told that their specimen may or may not eventually be used for DNA sequencing) and one of those samples is randomly and blindly selected as the source actually used for library construction. In this way, not only will the identity of the individual whose DNA is chosen not be known to the investigators, but that individual will also not be sure that s/he is the actual source. Return to Text
4 Although recontacting donors should not be possible, investigators will potentially want to be able to resample a donor's genome. Thus, at the time the initial specimen is obtained, in addition to making a clone library representing the donor's genome, it should also be used to prepare an additional aliquot of high molecular weight DNA for storage and a permanent cell line. Either resource could then be used as a source of the donor's genome in case additional DNA were needed or comparison with the results of the analysis of the cloned DNA were desired. Return to Text
5 There has been discussion in the scientific community about the sex of DNA donors. A library prepared from a female donor will contain DNA from the X chromosome in an amount equivalent to the autosomes, but will completely lack Y chromosomal DNA. Conversely, a library prepared from a male donor will contain Y DNA, but both X and Y DNA will only be present at half the frequency of the DNA from the other chromosomes. Scientifically, then, there are both advantages and disadvantages inherent in the use of either a male or a female donor. The question of the sex of the donor also involves the question of the use of somatic or germ line DNA to make libraries. For making libraries, useful amounts of germ line DNA can only be obtained from a male source (i.e., from sperm); it is not possible to obtain enough ova from a female donor to isolate germ line DNA for this purpose. Opinion is divided in the scientific community about whether germ line or somatic DNA should be used for large-scale sequencing. Somatic DNA is known to be rearranged, relative to germ line DNA, in certain regions (e.g., the immunoglobulin genes) and the possibility has been raised that other developmentally-based rearrangements may occur, although no example of the latter has been offered. While some believe that the sequence product should not contain any rearrangements of this sort, others consider this potential advantage of germ line DNA to be relatively minor in comparison to the need to have the X chromosome fully represented in sequencing efforts and prefer the use of somatic DNA. Return to Text
6 Individuals whose DNA was used for library construction (with the exception of those created from deceased or anonymous individuals) should be fully informed about the risks and benefits described above, should freely choose whether they would like their DNA to continue to be used for this purpose, and their decision should be documented. Return to Text