Abstracts

Rationale: The spectrum of sudden death disorders are unified by diagnoses eliminating all alternative causes of death. These disorders include Sudden Infant Death Syndrome (SIDS), Sudden Unexpected Death (SUD), and Sudden Unexpected Death in Epilepsy (SUDEP). Molecular analysis shows extensive patterns of personal variation in neuro-cardiorespiratory disease genes, limiting the utility of diagnostic gene testing for intervention. Our multiscale 'omics analysis shows known and suspected Sudden Death (SD) genes are alternatively spliced in the developing brain in age and region dependent patterns. This provides insight into the differential timing of sudden death disorders despite their shared molecular risk factors. Methods: We performed multiscale omics analysis combining a Next Generation Sequencing molecular autopsy of 8 individuals who died of SIDS, SUDEP, or SUD, with the RNAseq BrainSpan Developmental Transcriptome datasets. The Broad Institute's GATK best practices workflow was employed, followed by annotation using SnpEFF, SnpSift, and VEP to generate individual genetic variant profiles. These were then mapped to all known coding and non-coding transcripts and subjected to bioinformatics analysis using CADD, Polyphen2, SIFT, and PhastCons to assess pathogenic consequences. We undertook a targeted screen of the RNAseq datasets for 309 samples across 7 developmental time points (25 weeks post conception to >60 years of age) for 235 sudden death (SD) candidate genes, with splice variation and exon inclusion assessed for 16 distinct brain regions. Results: There was extensive personal variation across samples, regardless of age of death, with multiple variants identified within the 235 candidate SD genes (low 387; high 1209 variants), including ion channel, immune, serotonin receptor, and hypoxia response genes. Bioinformatics analysis identified between 64 and 110 potentially pathogenic exomic variants, including variants of unknown significance such as the SCN5A S524Y missense mutation previously implicated in Brugada syndrome. RNASeq data for 4487 exons was analyzed for 232 SD genes revealing different spatio-temporal expression patterns through development where highly complex and alternatively spliced genes (TTN) contribute to variant risk in a non-linear fashion compared to smaller single exon genes (CSTF2T, KCNE1L). Conclusions: This study offers unprecedented insight into the developmental relationships of the neuro-cardiorespiratory candidate genes implicated in the spectrum of sudden death disorders. For the first time, we have identified when, and in what form each gene product is expressed, and linked these patterns to personal risk to better inform preventative diagnostic screening. Funding: This work was funded in part by a CURE grant (Klassen Co-I) and CFI infrastructure to Klassen.