Diagnostic exome sequencing. Severe intellectual disability (ID) is unexplained in the vast majority of patients and is thought to be genetic. The genetics of intellectual disability has traditionally focused on the X chromosome, where more than 100 possibly causative genes for ID are located. But other, autosomal genes are also found in large number of cases. A recent study in the New England Journal of Medicine now reports on trio exome sequencing in patients with unexplained severe intellectual disability. The authors identify causative de novo events in a large proportion of patients. Interestingly, more than half of their patients had epilepsy.

Known knowns and unknown unknowns. This way of delineating the limits of knowledge pretty much summarizes the approach for diagnostic exome sequencing in severe intellectual disability, a disease with several hundreds of known candidate genes. Exome sequencing in trios can identify known ID genes (the known knowns) and a large number of genes that might be candidates. How large the number of the second group might be and how to tell causative genetic variants from genomic noise is not known (unknown unknowns). Depending on the filters used, every individual carries between 3-5 functionally relevant de novo mutations in his or her exome. As seen in autism, statistics on a group level can be slightly frustrating. The baseline number of de novo events in unaffected individuals makes it difficult to identify a slight excess in patients with disease. De Ligt and colleagues performed exome sequencing in 100 patients with unexplained ID and analyzed the data for de novo mutations.

Exome sequencing in severe ID. De Ligt and colleages have performed exome sequencing in 100 patients in severe ID, identifying de novo mutations in known ID genes and many candidate genes. Three of these novel genes are recurrent.

13% + x. The authors identified 13 patients with probably pathogenic de novo mutations in known ID genes, providing a good baseline for the yield of diagnostic exome sequencing in ID. With 87% of all targeted exons with a read depth of 10 or higher, it might be interesting to see to what extent this number might be increased when applying targeted arrays rather than exome sequencing. However, in contrast to targeted arrays, exome sequencing also offers the possibility to identify novel genes. The authors found potentially causative de novo variants in 22 patients. As it is difficult to tell causative mutations from random events, the authors focused on genes that were either already reported in the recent literature (DYNC1H1) or identified in a second cohort of > 700 patients with ID. The fact that only two patients could be identified in this cohort with a new mutations in GATAD2B and CTNNB1 despite analyzing up to 5 genes in all patients, tells us something about the genetic architecture of ID. Variants in each candidate gene are very rare. While these three additional genes can be considered pathogenic due to the recurrence in patients, the role of the other variants is unknown. Therefore, the overall yield of diagnostic exome sequencing is somewhere between 13% and 35%, and both a lower and upper limit can be delineated.

A de novo variant in West Syndrome. Extrapolating the frequency of de novo mutations in ID might be slightly biased since the cohort is highly selected. All patients were screened for various metabolic disorders and for CNVs. In addition, many patients had additional features and more than 50% of patients had epilepsy. Data mining the appendix to extract some more information on the phenotypes reveals the following preliminary information: Patient 21 had West Syndrome and a de novo mutation in MFAP3. Patient 91 had myoclonic epilepsy and a de novo mutation in EEF1A2. Patients 61 and 96 had absence epilepsy and de novo mutations in ACO1 and RBL2, respectively. Whether these mutations are causative is difficult to tell. However, they add to a growing body of data on de novo mutations in neurodevelopmental disorders. Which makes you wonder if anybody is actually keeping track of this?

What the authors did not find. The authors also looked at autosomal recessive mutations and did not find any. This is surprising, given the size of the cohort. Ever since the publications on the phenotypic spectrum of ADLH7A1, the gene causing classical vitamin B6 deficiency, I felt that at least some of the unexplained cases of autism, epilepsy and ID might be due to atypical presentations of neurometabolic disorders that escape detection through screening. This “hidden neurometabolic hypothesis” has not really found any supporting evidence yet.

The Cinderella effect. Autism genetics has led the way. Schizophrenia genetics and ID genetics have accomplished it. Only epilepsy is lagging behind with translating trio exome studies into publications. In a previous post and during several of my presentations, I have called this phenomenon the Cinderella effect, a reference to William G. Lennox. While everybody is slowly getting antsy waiting for the first publication to appear, this time lag is nothing to be ashamed of. It is a result of the phenotypic complexity of the epilepsies and the diversity of the field. And with dropping prices and efforts to obtain and analyze genetic data, this diversity will eventually pay off. In the same way that Cinderella needed a little bit of time step out of the shadow of her step sisters, genetics will eventually be our glass slipper.

Ingo Helbig

Child Neurology Fellow and epilepsy genetics researcher at the Children’s Hospital of Philadelphia (CHOP), USA and Department of Neuropediatrics, Kiel, Germany