16988
Modeling gene expression and rare sequence variation identifies genes and subnetworks underlying autism risk

Saturday, May 17, 2014: 1:55 PM
Marquis D (Marriott Marquis Atlanta)
K. Roeder1, L. Liu1, J. Lei1, S. Sanders2, J. Willsey2, M. W. State3, J. D. Buxbaum4 and B. Devlin5, (1)Statistics, Carnegie Mellon University, Pittsburgh, PA, (2)Yale University, New Haven, CT, (3)Psychiatry, UCSF, San Francisco, CA, (4)Seaver Autism Center for Research and Treatment, New York, NY, (5)University of Pittsburgh, Pittsburgh, PA
Background:  

De novo, loss-of-function (dnLoF) mutations occur twofold more often in autism spectrum disorder (ASD) probands than their unaffected siblings. Multiple, independent dnLoF mutations in the same gene implicate the gene in risk and hence provide a systematic, albeit arduous path forward for ASD genetics. Willsey et al. (2013) identified brain gene expression networks as meaningful for organization and inter-relationships of ASD genes; and identified the mid-fetal prefrontal and motor-somatosensory neocortex as the region and developmental periods in which these genes tend to coalesce to confer risk to ASD.

Objectives:  

Using Brainspan gene expression from this critical nexus and whole-exome sequencing data, we aim to accelerate the search for ASD risk genes and to obtain subnetworks of genes that promote neurobiological assessment of function.

Methods:  

We use a novel algorithm, DAWN, to model two kinds of data: rare variants from exome sequence; and gene co-expression in the mid-fetal prefrontal and motor-somatosensory neocortex. The algorithm casts the ensemble data as a Hidden Markov Random Field in which the graph structure is determined by gene co-expression. The algorithm combines these interrelationships with node-specific observations, namely gene identity, expression, genetic data, and its estimated effect on risk to identify risk genes.

Results:  

DAWN identifies novel genes that plausibly affect ASD risk. Validation experiments provide strong evidence that this set contains many true risk genes. Indeed, in the validation experiment DAWN was able to distinguish the genes that will accumulate new de novo loss of function mutations better than any existing method.    

Conclusions:  

Our results implicate neuritic outgrowth, arborization, guidance, and terminal specification of both axons and dendrites, on the basis of genes predicted to affect risk. Thus we hypothesize our results converge on mediation of coordinated neurite development and that risk for ASD arises from disorganized patterns of arborization in addition to the often-described synaptic dysfunction.