NEW YORK (GenomeWeb) – In a pair of papers, two sets of researchers characterized the mutational landscape of autism spectrum disorder.
In one study, researchers mapped single-nucleotide variation, indels, and copy-number variations involved in ASD — the Hospital for Sick Children's Steven Scherer and his colleagues uncovered 18 new candidate genes in that effort. In the second, Massachusetts General Hospital's Michael Talkowski and his colleagues focused on large structural variations in people with autism, identifying seven major classes of structural variants and 16 recurrent subclasses.
As they reported today in Nature Neuroscience, Scherer and his colleagues conducted whole-genome sequencing using two different platforms on 5,205 individuals from autism-affected families. The researchers sequenced their cohort to an average 40.4X depth and identified an average 73.8 de novo SNVs and 12.6 de novo indels or CNVs in each individual with autism.
By combining the data they generated with information from other large-scale studies, the researchers uncovered 54 putative autism risk genes. These genes were gauged to be likely mutation-intolerant by the ExAC database. By focusing on the X chromosomes of affected males, the researchers noted an additional seven genes with multiple loss-of-function mutations.
They also detected some 400 CNVs per genome.
Of the 61 autism-risk genes the researchers identified, 18 had not been reported previously. A number of these novel genes were uncovered in multiple families: mutations in PHF3, which encodes a PHD finger protein that regulates transcription, were found in four families, while MED13 was mutated in three families.
Scherer and his colleagues noted that about 60 percent of these 61 autism-risk genes were associated with syndromes or phenotypes, typically related to intellectual disability, in the Online Mendelian Inheritance in Man database. In functional analyses, they found that 80 percent of the risk genes were connected to gene networks often involved in synaptic transmission, transcriptional regulation, and RNA processing.
The researchers added that they've made their data available with associated clinical information as part of the MSSNG portal.
Meanwhile, Talkowski and his colleagues reported last night in Genome Biology that they sequenced 686 people with autism from the Simons Simplex Collection, as well as three unrelated people with neurodevelopmental disorders, to generate a genome-wide map of large structural variation. With their long-insert whole-genome sequencing approach, the researchers reached a mean physical coverage of 105X and a mean insert size of 3.5 kilobases.
They identified 11,735 SVs, about 17 percent of which were balanced or complex SVs (cxSVs). By comparing this SV map to ones previously generated, the investigators estimated that 38 percent of the SVs they detected hadn't been reported before. More specifically, they reported that nearly 94 percent of the cxSVs were novel.
Among these cxSVs, Talkowski and his colleagues uncovered 16 unique subclasses of recurrent and common cxSVs. Ten of these 16 subclasses were unbalanced inversions, and most cxSVs included at least one inverted segment. In addition, they noted that most cxSVs were intrachromosomal.
Rare SVs appeared to be more deleterious than common polymorphic SVs, they added. Among their cohort, they found that rare SVs were larger than common SVs, twice as likely to disrupt multiple classes of non-coding elements and one-and-a-half times as likely to lead to loss of function. Ten loci — especially ones with known roles in neurological disorders — were significantly enriched for rare SVs beyond expectations.
The investigators also found instances of extreme rearrangements, including of chromoanasynthesis and chromothripsis.
This SV map suggests a "previously underappreciated abundance and diversity of cxSV" in human disease, the team wrote.