NEW YORK (GenomeWeb) – Two new Nature studies by independent research groups catalogued and characterized the role that rare, protein-coding mutations or variants play in autism spectrum disorder.
For one of the studies, the Broad Institute's Mark Daly, Joseph Buxbaum from Mount Sinai's Icahn School of Medicine, and their colleagues did exome sequencing on 3,871 individuals with ASD.
By comparing protein-coding sequences in these affected individuals with exomes from almost 10,000 unaffected ancestry-matched controls or parental controls, the researchers identified a set of 22 autosomal genes that were significantly more likely to carry rare variants in the individuals with ASD.
Another 107 autosomal genes showed hints of being prone to rare mutations in the ASD group, members of that team reported, albeit at lower levels of statistical significance. Some 5 percent of the ASD cases considered carried de novo loss-of-function changes to one of the genes in that set, which is comprised of evolutionarily constrained genes showing lower-than-usual mutation rates overall.
Authors of that study discovered an over-representation of genes belonging to pathways related to synaptic function, transcription and splicing, and chromatin remodeling in the 107-gene set.
In another Nature study, a team led by researchers from Cold Spring Harbor Laboratory, the University of Washington, and the University of California at San Francisco applied exome sequencing to 2,517 families that included a single child with ASD. The approach yielded exome sequences for 2,508 affected children, parents from each of the families, and 1,911 unaffected siblings.
With the help of protein-coding sequence comparisons from children from the same family that did or did not have ASD, the team determined that 13 percent of the de novo missense mutations it identified were contributing to an estimated 12 percent of ASD risk. Meanwhile, 43 percent of the de novo missense mutations deemed "likely protein disrupting" appeared to account for another 9 percent of the ASD diagnoses.
These de novo mutations, along with de novo copy number changes appeared to have a hand in almost one-third of the complete set of simplex ASD cases considered for the study. Amongst females with ASD, that contribution was even higher, at around 45 percent.
Data from that study also pointed to overlapping gene sets affected by de novo mutations in affected females and affected males who scored poorly on intelligence tests — sets of genes that included chromatin modifying genes, genes implicated in schizophrenia and intellectual disability, and other pathways. In contrast, males with ASD who had higher intelligence test scores tended to carry de novo mutations that affected a largely distinct set of genes.
A Nature Communications study by investigators in Australia and the UK described dramatic genomic rearrangements detected in a significant subset of esophageal adenocarcinoma cases (EAC).
The team started by using whole-genome sequencing and array-based copy number profiling to assess fresh-frozen tumor and matched normal samples from 22 individuals with EAC.
Along with an over-representation of cytosine to thymine transitions in protein-coding sequences and thymine to guanine swaps in other parts of the genome, the analysis uncovered complicated structural rearrangements and clusters of highly mutated sequences in a subset of tumors.
In the discovery cohort, more than one-third of the EAC tumors showed signs of extensive rearrangements resembling chromothripsis. When the researchers expanded their analysis to include another 101 EAC tumors using SNP arrays, they saw similar sorts of chromothripsis-like rearrangements in another 32 tumors.
Through these and other analyses, the study's authors found evidence that "large-scale chromosomal changes delivered by chromothripsis and [BFB] provide a mechanism for tumor suppressor gene loss or oncogene amplification, which may then be clonally expanded due to selective advantage, as reported in other cancer types."
Investigators from the University of Texas Health Science Center, the University of Washington, Harvard School of Public Health, and Brigham and Women's Hospital teamed up to develop software that uses identify-by-patterns across the genome to discern relatedness between individuals and put together family pedigrees.
As the team reported in the American Journal of Human Genetics, the "Pedigree Reconstruction and Identification" (PRIMUS) approach scores and constructs non-consanguineous pedigrees that include up to third-degree relatives through analyses of genome-wide pairwise IBD.
In their proof-of-principle introduction of PRIMUS, the researchers first validated the software using thousands of simulated pedigrees, along with HapMap3 pedigrees, and more than 100 pedigrees considered in a clinical setting at the University of Washington's Center for Mendelian Genomics, demonstrating that it could accurately untangle pedigrees and even recognize unknown third-degree relationships in the HapMap data.
From there, they applied PRIMUS to array-based SNP data for nearly 1,900 individuals enrolled in a health study in Starr County, Texas. There, the software successfully assembled 203 family pedigrees using samples from 458 of the individuals.
"Because PRIMUS computationally verifies reported pedigrees by using genotype data and identifies and corrects inconsistencies, PRIMUS saves a significant amount of time and effort that would otherwise be spent on manual verification of pedigrees," the study's authors wrote. "Moreover, PRIMUS can reconstruct previously unknown pedigrees by using only genetic data, as demonstrated in the HapMap3 and Starr County data sets."
An international team used genome and transcriptome sequencing to characterize and compare the genomic architecture of clear cell renal cell carcinoma kidney cancer tumors in patients from four European countries.
As they reported in Nature Communications, the researchers used whole-genome sequencing, RNA sequencing, and array-based SNP profiling to assess tumor and matched normal samples from 94 individuals from Romania, Russia, the UK, and the Czech Republic who had clear cell renal cell carcinoma.
From that data, the team saw some 4,904 somatic mutations per tumor, on average, along with anywhere between zero and four-dozen copy number changes. The latter collection included a recurrent loss involving part of chromosome 3 that was picked up in 90 percent of the tumor samples. Chromosome X sequences were lost in more than 7 percent of samples.
The researchers also described unusual transcriptional profiles in the tumors, as well as recurrent alterations in genes coding for components of the PI3 kinase/mTOR signaling pathway and other pathways.
When they compared tumors from patients in different countries, meanwhile, the team saw an over-representation in nucleotide substitutions associated with exposure to aristolochic acid — carcinogenic compounds produced by European birthwort, wild ginger, and other plants.
"These results show that the processes underlying [clear cell renal cell carcinoma] tumorigenesis may vary in different populations and suggest that [aristolochic acid] may be an important [clear cell renal cell carcinoma] carcinogen in Romania, a finding with major public health implications," authors of the study explained.