ABSTRACT
For years, the genetics community has estimated the number of individual rare genetic diseases to be approximately 6,000-8,000. A commonly quoted derivation of this estimate is based on the simple addition of the number of phenotypic entries with and without confirmed molecular etiologies in the Online Mendelian Inheritance in Man (OMIM®). Here, we examine the validity of this estimation by mining the phenotypic entries in OMIM that are of likely or suspected Mendelian inheritance without a molecular cause (MIM number prefix "%" or "null"). Of the 3,204 unsolved phenotypic entries in OMIM, only two-thirds (2,034 entries) represented rare diseases. Of these, 8% were considered "well-established" based on their description in commonly used reference textbooks. We hypothesize based on the large proportion of entries that represent single families reported prior to 2011, that a number of the unsolved entries represent pathogenic variants in known genes. The novel gene discovery potential of these entries is therefore likely lower than originally thought. Given that the majority of the ~300 new disease-gene associations curated each year by OMIM were never associated with a "%" or "null" sign, the true scope of the rare disease atlas is likely much larger than previously anticipated.