Characterisation of the SMN1/2 locus using a highly specific variant caller on whole-genome sequence data from 500,000 individuals
摘要
Spinal muscular atrophy (SMA) is a rare autosomal recessive condition caused by biallelic loss of the SMN1 gene that results in severe lower motor neuron degeneration in childhood. Targeted pre-symptomatic treatments have been shown to dramatically improve outcomes, supporting the inclusion of SMA in newborn screening programmes. However, ~95% of infantile-onset cases are caused by homozygous deletions of exon 7-8 of the SMN1 gene, which is challenging to detect using short-read next-generation DNA sequencing technology due to the high sequence similarity between SMN1 and its nearby paralog SMN2. Here, we evaluate the performance of an SMN-specific variant caller in ~490,000 adults with whole-genome sequence data in UK Biobank. We also perform a phenome-wide association study (PheWAS) of 4782 diseases and traits for SMN1 deletion carriers. We show that the SMNCopyNumberCaller performs extremely well, identifying 8856 (1.8%) heterozygous carriers and just two (0.0004%) individuals with homozygous SMN1 deletions, of whom one had been diagnosed with SMN1-SMA. From the PheWAS, we found novel associations between SMN1/2 copy number and the level of several circulating proteins encoded or regulated by genes near the SMN locus, further validating the accuracy of the caller. Our results suggest that using a specialist variant caller can accurately determine SMN1/2 copy number from short-read DNA sequencing technology, achieving a very high specificity (~100%) for homozygous SMN1 deletions. The number of false positive results for SMN1-SMA is therefore likely to be extremely low when using short-read whole-genome sequencing at a population level.