Making Sense Of Nonsense : consequence of premature Stop mutations. S. Balasubramanian1, Y. Fu1, M. Pawashe1, M. Jin1, J. Liu1, D. MacArthur2, M. Gerstein1 1) MB & B, Yale University, New Haven, CT; 2) ATGU, Massachussets General Hospital, Boston, MA.
Loss-of-function variants (LOF) attract great clinical interest, as it is believed that most of them are potentially pathogenic. However, some LOF variants are also known to be beneficial. For example, LOF variants in PCSK9 lead to low LDL levels. Therefore, several pharmaceutical companies are actively pursuing the inhibition of PCSK9. Recent sequencing efforts demonstrate the presence of null variants in seemingly healthy people. Consequently, there is great interest in understanding putative LOF variants. While there are several methods for inferring the effect of nonsynonymous variants, there is no predictor for nonsense variants. Moreover, most methods predict pathogenicity of a variant without taking into consideration its genotype. Here, we present a method to infer the effect of LOF variants, primarily those due to premature Stop codons. We have developed ALOFT (Annotation of Loss-of-Function Transcripts), a pipeline to annotate putative LOF variants with a variety of functional and evolutionary features. Using these features, we developed a predictive model to classify nonsense variants into those that are benign, lead to recessive disease and those that lead to dominant disease. We applied this method to infer the effect of nonsense mutations in studies from the Center For Mendelian Genomics and correctly predict the mode of inheritance and pathogenicity of all of the truncating variants. We also validate our method by applying our classifier to four different autism studies. De-novo LOF SNPs have been implicated in autism. Our method shows that de-novo LOF events are significantly higher in autism cases versus controls. To our knowledge, this is the first method that predicts the impact of nonsense SNPs in the context of a diploid model, i.e. whether nonsense SNP will lead to recessive or dominant disease.
You may contact the first author (during and after the meeting) at