JMIR Publications

ABSTRACT

Background:

In the United States, there are 3 million people who have Autism Spectrum Disorder (ASD), and around 1 out of 59 children are diagnosed with ASD. People with ASD have characteristic social communication deficits and repetitive behaviors. The causes of the disorder remain unknown, however, in up to 25% of cases, a genetic cause can be identified. Detecting ASD as early as possible is desirable because early detection of ASD enables timely interventions on the children with ASD. Identification of ASD based on objective pathogenic mutation screening is a major first step towards early intervention and effective treatment of affected children.

Objective:

Recent investigation interrogated genomics data for detecting and treating autism disorder, in addition to the conventional clinical interview as a diagnostic test. Since deep neural networks performs better than shallow machine learning models on complex and high-dimensional data, in this paper, we sought to apply deep learning to genetic data obtained across thousands of simplex families at-risk for autism spectrum disorder to identify contributory mutations and create an advanced diagnostic classifier for autism screening.

Methods:

After preprocessing the genomics data from the Simons Simplex Collection, we extracted top ranking common variants that may be protective or pathogenic for autism based on Chi-Square test. A convolutional neural network-based diagnostic classifier is then designed using the identified significant common variants to predict autism. The performance is then compared with shallow machine learning-based classifiers and randomly selected common variants.

Results:

The selected contributory common variants are significantly enriched in chromosome X while chromosome Y is also discriminatory in determining identification as autistic or non-autistic of an individual. ARSD, MAGEB16 and MXRA5 genes had the largest effect in the contributory variants. As a result, screening algorithms were adapted to include these common variants. The deep learning model yielded an AUC of 0.955 and an accuracy of 88% for identifying autism from non-autism individuals. We demonstrated a significant improvement over standard autism screening tools.

Conclusions:

The common variants are informative for autism identification. These findings also suggest that deep learning process is a reliable method in distinguishing the diseased group from the control group based on the common variants of autism.