With this first post I want to briefly discuss a recently published paper, led by Alice Davidson (UCL Institute of Ophthalmology). This is a nice paper with two stories related to two different eye diseases, with one of them intriguing enough to motivate this post.
The main story: RP1L1 and retinitis pigmentosaThis paper started as a study of the genetics of retinitis pigmentosa (RP). As an aside, RP is an incredibly complex disease with a ridiculously long list of associated genes, each (or at least most) in a Mendelian type manner. In fact, it is probably fair to say that after the brain the eye must be the most complex human organ, at least based on the incredibly complex genetic architecture of these developmental traits.
So, in this study, exome sequencing in a RP patient identified a homozygous loss-of-function variant in a massive candidate gene (owing to homology with another RP gene). So it is quite certain that we have found yet again a novel gene for RP, which is the key finding of the paper.
The second story: occult macular dystrophyOwing to previous reports of association between RP1L1 variants and another eye disease called occult macular dystrophy (OCMD), Alice also scanned this gene in a cohort of 28 OCMD case. The results were convincing and consistent with previous work: 5 out of 28 harboured the same rare variant (pArg45Trp) which had been reported before in a Japanese study of 3 OCMD families (Akahori et al., 2010). In contrast the NHLBI exome sequencing dataset only reports one pArg45Trp call in more than 5,000 individuals. So there is no doubt that pArg45Trp is associated with OCMD. And we found other rare variants in this cohort, together explaining probably 9/28 cases.
Occult macular dystrophy is not a simple Mendelian diseaseSo far so good. But the story became rapidly more complex. Intriguingly, we found in the family of these probands 9 patients with the same RP1L1 clearly causal variants but unaffected (and old enough to have potentially developed the disease). So the RP1L1 variants are probably not fully penetrant (i.e. being a carrier does not mean one will develop the disease). As pointed out above, we also found 19/28 OCMD cases without any RP1L1 variants. SO RP1L1 is neither sufficient nor necessary to cause OCMD, in spite of the massive association.
A middle ground between complex and Mendelian?The emerging picture is that OCMD is half-way between a complex multi-factorial trait and a Mendelian one: there are very strongly associated variants in RP1L1, almost like a Mendelian trait, but this is not quite enough to explain the heritability that we observe. Nevertheless, the genetic architecture may still be simple enough to obtain in the future a more thorough, if not complete, dissection of the trait.
Now is that a common situation? I suspect it really depends on the disease field but it's definitely unusual for me. Cardiologists who work with diseases like hypertrophic cardiomyopathy are very familiar with these non fully penetrant variants and these complicated pedigrees that do not look fully Mendelian. But in my past experience I have mostly worked on Mendelian disorders with a very clear pattern of inheritance, or alternatively very complex multifactorial traits (like type 1 diabetes). It is somewhat rare for me to see such disorders with an intriguing architecture but perhaps simple enough to solve fully in the future, with the help of modern DNA sequencing technologies.