Response to Hamosh et al.

Main text

To the Editor: We are grateful that Hamosh et al. have engaged with our proposal regarding the naming of Mendelian genetic disorders.¹ Rather than our composing a tedious rebuttal of each of the six points they raise, we instead focus on a few major implications of their letter.

Hamosh et al. do not address the fundamental point raised in our proposal. The question we have asked is “what is a unitary and distinct genetic disorder”? We agree with Hamosh et al. that it is a bit of a mess. To address that challenge, we have made a fundamental assertion that the combination of an altered gene and a specific phenotype (the dyad) comprises a unitary and distinct disorder. That is, when considering genetic disorders, if either the phenotype is meaningfully distinct or the gene is distinct, the dyad is considered a distinct disease. This is an important conceptual advance that reflects two critical attributes of disease: the biology (etiology) and the clinical reality (phenotype).

Hamosh et al. reject the concept of incorporating etiology into a diagnostic descriptor. The gene is a powerful concept and must be incorporated into our conceptual models of disease and our diagnostic descriptors. It has been true for decades that “syndrome” has been used differently in our field compared to other medical applications of the term. We recognize that in some medical circles the term refers to a symptom complex (e.g., nephrotic syndrome) where the etiology may be unknown. But notably, the incorporation of etiology into the concept of “syndrome” as a disease designation was emphasized by two international working groups.^2,3 As Hamosh et al. note, this concept is also implicit and coded in OMIM. Instead of the dyadic descriptor that directly specifies the gene and the phenotype that we have proposed (e.g., PCYT2-related spastic paraplegia), OMIM specifies this as, for example, “spastic paraplegia 82.” What the dyadic proposal has done is to make explicit the gene as it is named (e.g., PCYT2) rather than requiring one to find the gene's OMIM code number, which is “82” for PCYT2. We do not find a cryptic numerical code for a gene to be useful and clinicians generally do not use these numbers given the difficulty of remembering such an arbitrary disease naming system. It is our view that the conceptual basis of incorporating etiology into a disease descriptor is well-established across many domains of medicine—it is hardly exceptional. It is puzzling that geneticists would recoil from including explicit specification of the gene in a disease descriptor.

As has been wryly said, “All models are wrong. Some are useful.”⁴ We humbly concede that the dyadic approach is, like all human cognitive models of pathophysiology, imperfect.⁵ At the same time, we find it to be the most useful of the available models of diagnosis. Our field has undergone dramatic advances since the latter half of the 20^th century. The astonishing advances in molecular biology demand that our concepts of disease must evolve.⁶ Although Hamosh et al. have raised some valid criticisms of our proposal, the most important part of their letter is what is missing—an alternative that is superior to the dyadic concept. After listing six issues that they have with the dyadic system, they conclude with only a desire to work on a naming approach—but have not advanced the debate in any constructive way by suggesting what that approach should be. Although we respect and accept criticisms of our approach, what we would most eagerly welcome would be a novel proposal for the designation of unitary and distinct genetic disorders that is superior to the dyadic approach that we have proposed. We look forward to working with our colleagues to improve the dyadic approach to mitigate the limitations that have been pointed out until such time as a superior proposal is advanced.