A new study in Science Translational Medicine,A method to delineate de novo missense variants across pathways prioritizes genes linked to autism,” applies a novel analytical method on a class of mutations identified in individuals with autism spectrum disorders (ASD) and predicts the severity of the disease’s complex symptoms.

The severity of autism symptoms falls on a spectrum, ranging from mild social deficits to severe cognitive impairment and disabilities. Compounding its complexity, ASD is a multifactorial disorder with environmental variables (such as pollutants) and multiple classes of genetic mutations playing contributing roles. The gamut of genetic studies in ASD suggest that mutations in hundreds of genes are involved in the development of the variable symptoms of autism.

Establishing conclusive links between alterations in the genetic code (genotype) and observable traits (phenotypes) are a difficult undertaking even in single-gene disorders. The genetic complexity inherent in complex disorders like ASD frustrates efforts at understanding how specific genotypes impact phenotypes and poses a major obstacle in identifying neuropsychiatric drug targets and developing safe and effective therapeutics.

Olivier Lichtarge, MD, PhD, professor of biochemistry and molecular biology at Baylor College of Medicine

Olivier Lichtarge, MD, PhD, professor of biochemistry and molecular biology at Baylor College of Medicine decided to investigate the relationship between autism symptoms and de novo missense variants. A de novo missense variant is a change in the protein coding sequence that alters the identity of a residue and that arises for the first time in a family member. Although individually rare, de novo mutations may constitute a significant heritable component of complex genetic diseases.

The researchers use an advanced computational prediction method called the “evolutionary action [EA] method” to analyze genetic data and de novo missense variants in 2,384 individuals with autism and 1,792 normal siblings who serve as controls.

“The immediate practical significance [of the EA method] is to show that the past evolutionary history of genome variations and divergences between species yield powerful additional information for interpreting human variations in complex diseases or traits. At a more fundamental level, the EA equation is born of a representation of the fitness landscape in which we can apply calculus to measure forces, displacements and energy potential just as we do normally in the physical world. In that sense, the use of EA over a large cohort of patients and controls unifies basic concepts of biology and physics, and of molecular evolution and population genetics,” says Lichtarge.

The biological effect of missense mutations on protein function can be scored using the EA equation that evaluates two terms that can be computed from evolutionary data: “sensitivity” and “perturbation size.” The EA score multiplies these two parameters to quantify the impact a mutation is likely to have had in evolution.

Sensitivity of a position in the sequence measures the degree of divergence of species (speciation event) when there is a change at that spot during evolution. For instance, a genetic change at a position that led to the separation of the dog and wolf—a small speciation event—would have lower sensitivity than a position in the sequence that when altered led to the separation of non-nucleated organisms (prokaryotes) from nucleated organisms (eukaryotes).

Perturbation size on the other hand, is a measure of the change in the biophysical properties of the amino-acid side-chain when an amino acid residue in the sequence is altered.

“The larger the EA [score] is, the more deleterious it is likely to be today in the current biological system of interest,” says Lichtarge.

The authors found missense variants in 398 genes that code for components in 23 biological pathways involved in development of neuronal axons, synaptic transmission, and neural development. In addition, the predicted fitness in patients with de novo and inherited missense mutations in candidate genes correlates with the IQ of individuals with ASD, even for new genes implicated in ASD.

Adopting the evolutionary action method, the authors detected those missense variants most likely to contribute to the disease phenotype in ASD and clarified the degree of their phenotypic impact. The implications of the scoring model that this study describes however, goes beyond ASD, as analogous approaches can be used to integrate missense variants across large groups of patients to identify genes contributing to a shared phenotype in other complex diseases.

“We can now ask whether the same score is also correlated with other important features of ASD, and whether it allows us to classify patients into different categories. Patient classifications are important to design more effective risk assessment and to evaluate therapeutic response of treatments. Some of the pathways and genes we identified may also provide new clues to candidate drug targets for further study,” says Lichtarge.