In humans, Strom et al. previously reported the heterozygous missense variant c.269A>G (p.Tyr90Cys) in ARL3 in a European-descent pedigree with non-syndromic retinitis pigmentosa.27 The variant, which was rare, appeared de novo and was predicted to be pathogenic, was confirmed as heterozygous in three affected individuals, and was transmitted in an autosomal-dominant fashion. A second allele was not identified, and mechanistic evaluation was not carried out. On the other hand, here we have identified bi-allelic ARL3 changes that fully segregate with a classical JBTS phenotype, including retinal changes. Thus, although the connection between the de novo ARL3 variant and retinitis pigmentosa remains unexplained, it seems that bi-allelic ARL3 deleterious variants are sufficient to cause JBTS. The involvement of ciliopathy-associated genes in non-syndromic retinitis pigmentosa has been well described, so it would be of interest for the affected individual reported by Strom et al. to be investigated for the possibility of a second deleterious allele in trans in ARL3. It is also possible that, as reported here, bi-allelic mutations in ARL3 give rise to an extended phenotype compared with its reported dominant phenotype. A growing number of genes are known to cause distinct phenotypes according to whether dominant or recessive variants are inherited. For retinitis pigmentosa, mutations (typically nonsense) in RP1 were initially described in an autosomal-dominant pattern,28 followed by autosomal-recessive (homozygous missense) variants.29 For Gillespie syndrome, a form of non-progressive cerebellar ataxia, both bi-allelic and mono-allelic mutations in ITPR1 (MIM: 147265) have been reported,30 and the single heterozygous mutations were thought to exert a dominant-negative effect. In addition, variants in genes known only to be related to autosomal-dominant disease have been found in association with recessive mutations, both where the phenotypes are similar but more severe (ACTG2-related visceral myopathy [MIM: 102545]) and where distinctly different phenotypes have been observed (FBN2-related myopathy [MIM: 612570] and CSF1R-related brain malformation [MIM: 164770]).31