• Genetics
• metabolic disorders

In a seminal article in 2007, Ropers presented new perspectives for the elucidation of genetic disorders.1 He states what clinical geneticists have known for years—that Mendelian disorders, and in particular, autosomal recessive disorders, deserve more attention. Online Mendelian Inheritance in Man (OMIM) indicates that there are 1638 autosomal phenotypes for which the molecular basis is unknown. Historically, such disorders have been neglected, both because of their rarity and because of the difficulty in identifying the underlying genes and mutations.

Recently, exome capture and sequencing has been shown to be a feasible approach to identifying disease mutations,2–4 in theory using only a single patient. Thus a systematic programme to discover all of these genes is an appropriate and timely objective.

Those of us working in the area of clinical research often give lip service to the concept that we are working for the direct benefit of patients, when in fact our goals are much more basic. The value of gene discovery for Mendelian disorders is that the results can be used immediately—if not for therapy, then at least for diagnosis, prognosis and genetic counselling. Although it is true that gene discovery often leads to a better understanding of pathogenesis and therapy, our short-term goal should be gene discovery for the sake of better diagnosis and genetic counselling.

In the ‘RaDiCAL’ (Rare Disease Consortium for Autosomal Loci) based at McGill, we concentrate our efforts on putative autosomal recessive diseases and start wherever possible with the analysis of a single proband. The reason for giving priority to autosomal recessive conditions is obvious. In taking a whole exome (or whole genome approach), the analysis will yield far fewer false positive hits if one is looking for either a homozygous deleterious mutation or two deleterious mutations in a single gene. With autosomal dominant disorders, only one mutation is required for a clinical phenotype, and analysis will be complicated by the large number of candidate heterozygous putative mutations generated by a genome-wide search.

Our approach will be to collect a single clinically well-described proband for each autosomal recessive disease for which the gene is not yet known. At the onset, there will be no attempt to collect families or additional patients with the disease. The aim is to collect one proband for each disease. Speed and the construction of a morbid map is the ultimate goal of RaDiCAL. The limiting factor in this approach is not the molecular biology but rather the clinical genetics. If the clinical diagnosis is incorrect, gene assignment will be incorrect. Because of the possibility of genetic heterogeneity within a presumed single Mendelian disease, all assignments will have to be independently validated in additional patients. In addition, the diseases to be studied will not be prioritised by the nature of the clinical condition, but rather by the quality of the characterisation of the proband.

Once probands have been identified, they will be prioritised on the basis of consanguinity, the availability of other affected and unaffected family members, and by the feasibility of easy functional validation.

We have been pleasantly surprised how, in many cases, exome sequencing has led to the appearance of a very limited number of candidate genes for the autosomal conditions we have looked at to date. The question that has troubled us a great deal is what is the threshold and mechanism of publication—or perhaps more correctly—getting the information quickly to the general community for validation and translation to medical genetics practice. There may be many mis-assignments, but getting the information out will allow rapid testing and confirmation of positive candidates. This approach should allow the discovery of the majority of causal genes for autosomal recessive conditions in the most rapid, if not the most elegant, manner.

Because of the rarity of many of the conditions and the absence of a priori functional assays, validation of the proposed candidate genes can best be carried out by clinicians in different centres worldwide. Resequencing of a candidate gene is a trivial pursuit and should be possible at any university centre.

What are some of the challenges to the rapid discovery of all causal genes for autosomal disorders? Clinicians will be reluctant to share their valuable clinical material if the seminal nature of their contribution to discovery is not recognised. Often the clinicians are relegated to middle authorship, when it is their accurate characterisation of the proband that makes large throughput of many conditions possible. We propose a new category of authorship that gives appropriate attribution to this role. Perhaps one solution could be co-first and co-last authorships shared by lead clinicians and lead molecular biologists, where appropriate. Perhaps a new classification of authorship will be recognised for gene discovery based on the clinical description of a single patient.

Another challenge may be the need for a generic consent in situations where limited numbers of patients are used for the study of large numbers of different diseases. It is not practical to have an individual consent form and IRB for each disease. For some of these conditions, the most appropriate patients may have died and cannot give their consent for the use of stored DNA or cell lines. In other instances, the probands may be infants or minors or have cognitive impairment.

A challenge of exome sequencing is that data may be found on mutations in known disease-causing genes that are not the subject of the study itself. Our preference is not to look specifically at known disease-causing genes that are not candidates by exome capture and sequencing for the disease being studied, and to not return the complete data on exome or whole genome sequencing to patients or their physicians. In effect, these will be looked at only insofar as necessary to eliminate moonlighting functions for known genes. In addition, we do not propose to post data in situations where no candidate genes stand out. Although we understand that these data may be useful, issues on confidentially and access to data on rare patients may complicate the underlying goal of rapid gene discovery.

Also, what becomes the threshold for publication in a situation where the real goal is putting forward candidate genes? Candidates that have been validated in other patients, and for which there are functional studies, will continue to be published using conventional peer review in established journals. But what about those situations where there are no supporting linkage data and no functional information. Perhaps, one or two genetics journals and their reviewers will recognise that value of publishing these candidates and perhaps develop specific sections to allow them to be seen and validated or excluded by other investigators. What about those situations where there are no obvious candidates, too many mutations and no validation? Is there a value of having a centralised database that lists what entities have been looked at with no results of interest? Indeed, a centralised database that simply posted candidate genes for the different conditions would speed up gene discovery. But in the current environment, who would set up and maintain such a database?

We have been challenged that no journal would want to publish ‘half-baked’ and possibly incorrect results, and that researchers would be unwilling to share their unpublished results! We have been told that altruism and the benefit of patients will not drive research if there is no money or prestige in the equation. Our challenge to the medical genetics community is to seize the opportunity, defy the sceptics and move forward.