ASD Subtype Study: Main Analysis and Figure Code

R and Python scripts for phenotype imputation, ASD subtyping, de novo variant annotation, genetic burden analysis, enrichment analysis, experimental candidate prioritization, and figure generation for the ASD subtype study.

Analysis

1. Phenotype imputation

• Impute phenotype matrices for discovery and replication cohorts from partially observed clinical measures.
• Generate harmonized phenotype tables used for downstream clustering and subtype analyses.

2. K-means subtyping

• Derive ASD subtype assignments from imputed phenotype data.
• Evaluate clustering structure across discovery, replication, and sensitivity settings.

3. DNV annotation

• Annotate de novo variants and classify splicing-disruption mechanisms.
• Assign coding severity, predicted protein impact, NMD status, and ASO amenability tiers.

4. Genetic burden analysis

• Compare subtype-stratified burden of DNVs, polygenic scores, and composite scores.
• Generate burden summary tables for subtype-to-subtype and control-to-subtype comparisons.

5. SDV burden by subtype

• Summarize severe splicing-disrupting variants across ASD subtypes.
• Produce carrier-level and gene-level tables for downstream visualization.

6. Gene and cell-type enrichment

• Perform gene-set, cell-type, and related enrichment analyses.
• Summarize subtype-linked genes in the context of known ASD gene resources and expression references.

7. Experimental candidate selection

• Prioritize ASD-associated splice-disrupting variants for downstream follow-up.
• Integrate annotation, severity, constraint, and subtype information into a candidate pool.

Figures

• Figure 1: Phenotype-based ASD subtype overview, including clustering structure and subtype profiles.
• Figure 2: Subtype differences in de novo variant burden.
• Figure 3: Detailed characterization of splicing-disrupting variant burden and carrier-level subtype patterns.
• Figure 4: Polygenic score burden, composite score comparisons, and subtype-level quantitative contrasts.
• Figure 5: Gene-level, pathway-level, and cell-type enrichment results linked to subtype-associated variation.
• Figure 6: ASO amenability landscape and prioritized experimental candidate summary.
• Extended Data Fig 1-10: Supporting analyses for imputation, clustering robustness, burden patterns, and gene-level examples.
• Supplementary Fig 1-5: Additional burden, score, and subtype comparison visualizations.