NEW YORK (GenomeWeb News) – The human genome shows a wide range of mutation rates along its expanse, and researchers led by Jonathan Sebat, an associate professor at the University of California, San Diego, reported in Cell today that genes associated with autism-spectrum disorders and other diseases are often located in regions of hypermutability.
This, the researchers said, indicates that hypermutability regions influence genetic variation patterns and disease risk.
"Some disease-related genes are gluttons for punishment," said Sebat, the paper's senior author, in a statement. "Despite the fact that these genes are important for normal human development, they appear to be getting hammered with mutations."
To explore such germline mutation patterns, the researchers sequenced the whole genomes of 10 monozygotic twins, who were concordant for an autism-spectrum disorder, and their parents. Drawing on the genomes of twins who both had autism-spectrum disorders offered two advantages, the researchers said, namely that they could more easily connect germline mutations to disease risk and that they could better tease out germline mutations from somatic mutations.
Additionally, to account for the effects of paternal age, half the twin pairs included in the analysis had younger fathers and half had older fathers.
The samples were sequenced at BGI using the Illumina HiSeq platform to a depth of 40x. From this, the investigators identified a set of 581 de novo mutations affecting 34 genes, 29 of which are protein-coding genes and five of which are noncoding genes.
That, the researchers said, suggests an average human genome-wide mutation rate of 1x10-8 per generation — a figure that is about two-fold lower than previous estimates, but that is in line with estimates from studies recently published in Nature and Nature Genetics.
Additionally, the germline de novo mutations were found to be distributed nonrandomly throughout the genome. Indeed, many were found to be contained within clusters, which the researchers defined as two or more de novo mutations within 100 kilobases. Some regions, the investigators said, exhibited as much as 100-fold higher mutation rates than other regions of the genome.
A number of intrinsic factors, such as GC content, nucleosome occupancy, and chromatin structure, among others, affected mutability levels, the researchers reported.
By combining those intrinsic features that affect mutability, the researchers developed a predictive model to gauge a region's mutability index, which they reported could explain about 90 percent of the difference in mutation rates seen across the genome. The researchers confirmed their model using an independent dataset.
Meanwhile, another mutation pattern the researchers uncovered indicated a U-shaped relationship between mutability and evolutionary conservation.
"Paradoxically, some of the mostly highly mutable sequences in the genome are in fact highly conserved," the researchers noted. They suggested three possible theories to explain this seeming contradiction: one, that hypermutability itself could be under selection; two, that functional or conserved elements may have derived from ancient hotspots; and, three, that hotspots arise due to some DNA repair processes being tied to gene regulation and transcription.
Then, with the predictive model, the researchers pinpointed so-called hotspots of mutability — as well as warm, cool, and cold spots — in the genome.
Regions of hypermutability, the investigators found, are associated with disease genes. Additionally, from a literature search, the researchers noted that genes expressed in the brain tended to be in areas of higher mutability, and a number of autism-spectrum disorder-associated genes, including NRXN1, AUTS2, GABRB3, SHANK2, and KCNMA1, were found to contain some of the top mutable exons in the exome.
"Our findings provide some insights into the underlying basis of autism — that, surprisingly, the genome is not shy about tinkering with its important genes," said Sebat. "To the contrary, disease-causing genes tend to be hypermutable."
The de novo mutations the researchers found in the twin pairs that affected protein-coding genes were also found to be associated with autism-spectrum disorders in separate exome sequencing samples, and those genes include "intriguing candidates" for autism risk genes, especially GPR98 and KIRREL3.