Sleat_2016_Gene_593_284

The neuronal ceroid lipofuscinoses (NCLs) are a group of fatal, mostly recessive neurodegenerative lysosomal storage diseases. While clinically similar, they are genetically distinct and result from mutations in at least twelve different genes. Estimates of NCL incidence range from 0.6 to 14 per 100,000 live births but vary widely between populations and are influenced by whether patients are classified based upon clinical or genetic criteria. We investigated mutations in twelve NCL genes in ~61,000 individuals represented in the Exome Aggregation Consortium (ExAC) whole exome sequencing database. Variants were extracted from ExAC and pathogenic alleles were differentiated from neutral polymorphisms using annotated variant databases and missense mutation prediction tools. Carrier frequency was dependent on ethnicity, with the highest (1/75) observed for PPT1 in the Finnish. When data are adjusted for ethnic diversity within the USA, PPT1, TPP1 and CLN3 carrier frequencies were found to be the highest of the NCLs, each at ~1/500. Carrier frequencies calculated from ExAC correlated well with incidence estimated from numbers of living NCL patients in the US. In addition, the analysis identified numerous variants that are annotated as pathogenic in public repositories but have a predicted frequency that is not consistent with patient studies. These variants appear to be neutral polymorphisms that are reported as pathogenic without validation. Based upon literature reports, such alleles may be annotated in public databases as pathogenic and this propagates errors that can have clinical consequences.