De novo structural variants in autism spectrum disorder disrupt distal regulatory interactions of neuronal genes

Three-dimensional genome organization plays a critical role in gene regulation, and disruptions can lead to developmental disorders by altering the contact between genes and their distal regulatory elements. Structural variants (SVs) can disturb local genome organization, such as the merging of topologically associating domains upon boundary deletion. Testing large numbers of SVs experimentally for their effects on chromatin structure and gene expression is time and cost prohibitive. To address this, we propose a computational approach to predict SV impacts on genome folding, which can help prioritize causal hypotheses for functional testing. We developed a weighted scoring method that measures chromatin contact changes specifically affecting regions of interest, such as regulatory elements or promoters, and implemented it in the SuPreMo-Akita software (Gjoni and Pollard 2024). With this tool, we ranked hundreds of de novo SVs (dnSVs) from autism spectrum disorder (ASD) individuals and their unaffected siblings based on predicted disruptions to nearby neuronal regulatory interactions. This revealed that putative cisregulatory element interactions (CREints) are more disrupted by dnSVs from ASD probands versus unaffected siblings. We prioritized candidate variants that disrupt ASD CREints and validated our top-ranked locus using isogenic excitatory neurons with and without the dnSV, confirming accurate predictions of disrupted chromatin contacts. This study establishes disrupted genome folding as a potential genetic mechanism in ASD and provides a general strategy for prioritizing variants predicted to disrupt regulatory interactions across tissues.

Preprint on bioRxiv

---

In this repo:

1. Simons Simplex Collection dnSVs
2. Scoring SSC dnSVs with SuPreMo-Akita
3. Scoring SSC dnSVs with CREint weights
4. Filtering SSC dnSVs using selection criteria
5. HiC data analysis
6. RNAseq data analysis

1. Simons Simplex Collection dnSVs

De novo structural variants used in this study are from Simons Simplex Collection (SSC), as a part of Simons Foundation Autism Research Initiative (SFARI). Belyeu et al 2021 called dnSVs in hg38 using alignment-based, short-read WGS, and we pulled them from their Supplementary Table 1 into data/SFARI_SSC_dnSVs.csv.

2. Scoring SSC dnSVs with SuPreMo-Akita

We installed SuPreMo-Akita and cloned the SuPreMo repo into this repo.

We scored SSC dnSVs:

python SuPreMo/scripts/SuPreMo.py variant_scoring/supremo-akita_input/dnSVs_for_SuPreMo.txt \
--get_Akita_scores \
--dir variant_scoring/supremo-akita_output \
--file dnSV \
--fa SuPreMo/data/hg38.fa

Our steps to process the input and output files are in variant_scoring/scoring_dnSVs.ipynb.

3. Scoring SSC dnSVs with CREint weights

CREints were processed from Song et. al. 2020 excitatory neuron H3K4me3 PLACseq data, pulled from NeMO into data/eN.MAPS.peaks.txt.

We scored SSC dnSVs near CREints with weighted scoring:

python SuPreMo/scripts/SuPreMo.py variant_scoring/supremo-akita_input_weighted/EP_for_SuPreMo.txt
--get_Akita_scores
--shifts_file variant_scoring/supremo-akita_input_weighted/EP_for_SuPreMo_shifts.txt
--roi variant_scoring/supremo-akita_input_weighted/EP_for_SuPreMo_weights.txt
--roi_scales 10 1000000
--dir variant_scoring/supremo-akita_output_weighted
--file EP
--fa SuPreMo/data/hg38.fa

Our steps to process the input and output files are in variant_scoring/scoring_dnSVs.ipynb.

4. Filtering SSC dnSVs using selection criteria

We defined a set of criteria to prioritize variants that are likely to be causal and can feasibly be tested in excitatory neuronal cells. The criteria and the variants that pass them (data for Figure S3A-B) are in variant_prioritization/prioritizing_dnSVs.ipynb.

5. HiC data analysis

To process HiC fastq files into mcool files, we use the 4DN pipeline. The code in HiCanalyses/hic_analysis.sh has beed adapted from the 4DN HiC Docker GitHub repo. We used pacakge versions shown in HiCanalyses/hic_analyses.yml and ran:

hic_analyses.sh <nthreads>, <genome_index>, <chrom_sizes>, <fastq1_rep1>, <fastq2_rep1>, <fastq1_rep2>, <fastq2_rep2>, <prefix_rep1>, <prefix_rep2>, <prefix>, <outdir>, <hic_analysis_path>

Raw and analyzed HiC data can be found in GEO (accession number GSE281283).

6. RNAseq data analysis

The differential gene expression and Gene Ontology enrichment analysis can be visualized in DEXanalysis.html. (Note: HTMLs need to be downloaded from github and then opened). All package versions are contained within and the associated R notebook can be used to re-run the code. The normalized counts, differential gene lists, enriched go terms, and figures can all be found in rnaseq/dex_output/.

The required inputs are located in rnaseq/input_data/ and include:

• /star -- contains the gene count matrix output from STAR v2.7.11b following the alignment to GRCh38.112 in gene annotation mode.
• /bamqc -- contains the quality control metrics generated using Picard v3.1.1
• multiqc_report.html -- coallates quality metrics from programs above and fastqc

Raw and analyzed RNAseq data can be found in GEO (accession number GSE281327).

---

For any questions and/or feedback, please reach out to katie.gjoni at ucsf dot edu