1.  The Academy of Medical Sciences represents many of the UK's most distinguished researchers in medicine and biology.

  2.  Fellows of the Academy are associated with many different types of genetic database, within the definition of "collections of genetic sequence information, or of human tissue from which such information might be derived". Time has not permitted a full survey of members' involvement, and details of individual projects will presumably be supplied by the relevant funding agencies.

  3.  This response therefore concentrates on the general types of collections of which we are aware, and the purposes and constraints on research involving these databases.

4.  TYPES OF GENETIC COLLECTIONS

  Genetic collections may be derived from research, or from clinical practice.

  4.1  Clinically derived collections include:

  4.1.1  Every molecular diagnostic laboratory will have banked DNA from patients who have been clinically tested for any of a wide variety of genetic disorders. These include:

—  Samples from individuals known to be affected with a genetic condition, where the present state of scientific knowledge or service development does not allow molecular diagnosis. When such diagnosis becomes available, other family members may want testing to see if they have inherited the mutation. Without material from an affected person, it is impossible to interpret negative mutation studies.

—  Samples from individuals with undiagnosed abnormalities (including babies with congenital abnormalities), stored in the hope that future research may allow a definitive diagnosis and hence accurate counselling for family members. Many of these samples are stored from individuals who are no longer alive.

  4.1.2  The majority of pathology laboratories (such as histopathology and haematology) will retain portions of tissue taken for diagnostic purposes. These samples are retained predominantly in order to allow subsequent confirmation or refinement of diagnosis, for the benefit of the patient or the patient's family, as new techniques evolve. However, they also form a potentially valuable research resource. Archived tumour samples are particularly important for genetic studies. These tissue samples can generate genetic information.

  4.1.3  Guthrie cards—regional neonatal screening programmes receive Guthrie cards, with a number of drops of blood on them, from virtually every newborn child in the United Kingdom. Some of these samples are stored for protracted periods of time; regrettably, long term storage has become less frequent over time. Although limited in quantity of material per person, these form a potentially very large source of DNA samples for some types of research.

  All these collections are funded as part of the NHS service, are linkable to NHS notes on the individual, with NHS service arrangements for confidentiality and access. Such collections are not subject to research ethics approval. Research use of these samples is (and should be) subject to search ethics review and approval.

4.2  Research derived databases

  These may be divided into collections of information and material relating to individual diseases, and multi-purpose collections (often populations collected as epidemiological cohorts). Large numbers of both types of study have been carried out in the past, and material from many of them is still stored. These studies will normally have been conducted under protocols agreed by research ethics committees. The samples often come from a mixture of two sources. The researcher will seek out suitable subjects and ask them to take part. Service departments will also have suitable material from subjects previously seen in clinic, stored with appropriate consent for future research use.

5.  PURPOSES OF GENETIC DATABASES

  5.1  The samples collected for clinical purposes, and in accordance with the standards required for accreditation of clinical laboratories, may be used for diagnosis, mutation analysis, and counselling of families with genetic diseases. Because they are carefully characterised, and can be associated with clinical information, they also have potential as research.

  5.2  Research collections have been assembled for a variety of purposes, in general with the approval of a research ethics committee, and often with external funding. However, where clinical samples have been used for research purposes, or have been secondarily assembled into research collections, issues of consent may be less clear. Where samples have been specifically collected by research laboratories, for research purposes, they may secondarily give rise to clinically relevant information. It is widely accepted that such information should only with great caution be fed back to patients or used for clinical purposes, unless it has been independently confirmed to the operating standards accepted in clinical diagnostic laboratories. Although this is widely accepted we are not aware that formal guidelines are in place to ensure uniformity of practice in this area.

  5.3  The great majority of such studies are intended to clarify the relationship between variation in human health, development, or behaviour with specific genetic variation. This is an area of very significant promise for the future, in terms of understanding the biological basis of medical conditions, from which may flow new diagnostic and therapeutic modalities.

  5.4  Most of these research studies are funded by external research funding organisations, such as the MRC and medical research charities.

  5.5  We do not consider that there are at present any alternative practicable ways of unravelling the biology of human health and disease, other than by studies correlating phenotype with genetic variation in one way or another.

  5.6  Our Fellows anticipate that research of this nature will expand in the future. They anticipate that current early exploratory studies will allow more focussed questions to be addressed particularly in relation to interactions between environmental factors such as diet or exercise, or toxic compounds in the environment or drugs, and particular genotypes.

  5.7  It is also widely expected that technical capabilities for genotyping will become very, very much improved, and that costs of genotyping will come down dramatically over the next five years.

  6.  Several Fellows pointed to the important opportunities raised by the MRC and Wellcome calls for new case-control and cohort studies. It is expected that these will be carried out in accordance with the MRC's recently issued ethical guidelines.

  7.  A number of potential constraints on this type of research either exist or can be anticipated. The costs of recruiting participants, collecting data and undertaking large numbers of genetic typings is considerable. The technology for preserving and maintaining an adequate supply of DNA for very large numbers of typings is as yet imperfect.

  The collection of clinical data in large prospective studies would be considerably enhanced if more effective electronic records were available in the NHS. This would entail developing mechanisms for obtaining appropriate patient consent. Without attention to these matters the value of the NHS as a research base, for the benefit of the health and economy of the country, will be greatly diminished.

  8.  It is of the utmost importance that the reasonable and appropriate concerns about privacy and confidentiality and appropriate use of genetic information should be openly discussed and resolved, to allow research for the benefit of society to go forward under proper control. While fully recognising the importance of public assent and individual consent to studies of this sort, there is some concern that it may be retarded by excessive conservatism on the part of some ethics review bodies. Fellows have experiences of Research Ethics Committees being extremely anxious about relatively straightforward genetic linkage studies on clinically derived samples because of a confusion between research and diagnostic tests, and the belief that research studies are likely to generate predictive health information which is only infrequently the case—clinical samples generally already have diagnostic information attached to them. An exaggerated perception of the extent of "genetic determinism" leads to undue fear about genetic studies. Many of the studies under consideration must deal with identifiable samples because of the need to correlate laboratory and genetic information, but once correlated the data are analysed and published in aggregate anonymised format and mechanisms for ensuring appropriate individual confidentiality are not excessively hard to devise. Research based on archived and anonymous samples is exceedingly unlikely to lead to any form of harm to individuals, and consent procedures should be set at levels appropriate to the risk of harm. This research will often require very large numbers, and will therefore be facilitated by access to diagnostic databases such as cancer registers. Considered over-regulation of privacy issues may damage these, without any corresponding gain to the public.

  9.  The purpose of these studies is to generate information relevant to health. Turning this basic research into usable health products will require commercial involvement. This should generally be of an open, non-exclusive nature, and must be fully explained to research subjects as part of the consent procedure before participation.

  This response was prepared on behalf of the Academy by Professor Martin Bobrow, FMedSci, Professor of Medical Genetics at the University of Cambridge, with contributions from Professor Patricia Jacobs, FMedSci, Director, Wessex Regional Genetics Laboratory, Salisbury, Professor David Porteous, FMedSci, Western General Hospital, Edinburgh, Professor Andrew Read, FMedSci, St Mary's Hospital Manchester, and Professor Ellen Solomon, FMedSci, Head, Division of Medical & Molecular Genetics, Guy's, Thomas' School of Medicine, London.