Unsupervised Dimensionality Reduction Techniques for the Assessment of ASD Biomarkers

Area of Science:

• Neuroscience
• Genetics
• Developmental Psychology

Background:

• Autism Spectrum Disorder (ASD) presents with diverse social, cognitive, and behavioral differences.
• Etiology of ASD involves complex genetic and environmental interactions, with unclear causal factors.
• A significant sex disparity exists in ASD diagnoses, suggesting sex-specific biological underpinnings.

Purpose of the Study:

• To develop integrative models for exploring gene expression effects on ASD-related traits.
• To investigate sex-specific biological influences in Autism Spectrum Disorder.
• To enhance the understanding of genetic and phenotypic heterogeneity in ASD, particularly concerning sex differences.

Main Methods:

• Utilizing principal component analysis (PCA) integrated with supervised machine learning algorithms.
• Analyzing multidimensional datasets from neuroimaging techniques and transcriptomics.
• Leveraging advanced data analysis methods to address challenges in evaluating complex ASD data.

Main Results:

• The methodology aims to navigate complex data spaces inherent in ASD research.
• The approach is designed to better contextualize genetic and phenotypic heterogeneity.
• This research seeks to uncover environmental and genetic interactions influencing ASD.

Conclusions:

• The developed methodology offers a robust approach to analyzing complex ASD datasets.
• This research paves the way for understanding sex differences in ASD.
• The findings are expected to contribute to the development of tailored interventions for Autism Spectrum Disorder.