NEW YORK (GenomeWeb) – A search for structural variants involved in autism spectrum disorder has uncovered a slew of large, de novo SVs affecting coding regions of genes in individuals with the neurodevelopmental condition.
"Children with autism do not have more mutations overall," Jonathan Sebat, a University of California, San Diego psychiatry researcher and the study's senior author, said in a statement, "but their mutations are more likely to disrupt genes involved in brain development."
Sebat and his colleagues did whole-genome sequencing on 235 individuals — 71 with ASD, 26 unaffected siblings, and parents when available. With the help of a SV-calling pipeline that incorporates several variant detection approaches, they tallied up tens of thousands of SV loci in the genomes, including larger de novo SVs occurring in conjunction with substitutions and/or small insertions and deletions in a single individual.
Overall, the team reported today in the American Journal of Human Genetics, around 20 percent of individuals with or without ASD had de novo structural mutations in their genomes, though SVs seemed to more frequently disrupt genes in those affected with ASD.
The team began by having whole-genome sequencing done on 246 individuals, including individuals from 11 identical twin pairs, at an Illumina Fast Track service laboratory or in collaboration with investigators at BGI. All of the samples were sequenced on Illumina HiSeq instruments to depths of around 40-fold coverage, on average. After excluding data for one twin per pair, the team was left with sequence information on the aforementioned 235 individuals.
The team then analyzed these genomes with a strategy that brought together three different methods for computationally detecting structural changes, uncovering more than 1.2 million suspected deletions, duplications, inversions, or mobile element insertions affecting more than 29,700 sites in the genome.
These SVs fell into 11 classes, including complex SVs — non-tandem duplications, duplication-inversion-duplications, and the like — that turned up some 251 times per person, on average.
To evaluate the role of de novo SVs in ASD, the team first attempted to weed out false positive de novo SVs using several features of the sequence data, and did array-based validation on potential de novo SVs.
All told, that search led to 19 validated de novo SVs affecting 19.7 percent of individuals with ASD and 19.2 percent of those without. Where the researchers did see pronounced differences between cases and controls, though, was in the size and anticipated effects of the de novo SVs: in individuals with ASD, the de novo SVs spanned a median of 10,900 bases compared to 1,200 bases in controls and were more likely to fall in the protein-coding portions of genes.
The team looked at the various SVs in more detail, identifying some of the genes altered by SVs in individuals with ASD as well as the structural variant combinations that tended to turn up together.
"[A] genetic contribution of de novo SVs to ASD is evident not from an elevated frequency of genomic rearrangement," Sebat and co-authors wrote, "but instead from the greater proportion of new mutations that disrupt genes."