Open analysis and pathway-level multi-omics integration for ASD: GWAS-derived genetic priors, pediatric proteomics anchors, and gut microbiome signatures.

1. Overview

Scientific rationale & importance.

Autism spectrum disorder (ASD) prevalence among 8-year-old children in the 2022 ADDM surveillance year was reported as 32.2 per 1,000 (~1 in 31), underscoring the need for scalable, earlier risk stratification and mechanistic insight (Shaw et al., 2025). Despite abundant omics datasets, cross-study biological signal is often underexploited due to heterogeneity in phenotyping, preprocessing pipelines, and batch effects. This proposal targets high-risk/high-reward but feasible proof-of-concept (POC) research by re-analyzing and integrating existing public datasets to test whether ASD-related signals converge at the pathway level across:

genomics priors derived from ASD GWAS summary statistics
pediatric proteomics datasets as biological anchors
gut microbiome 16S profiles from a public pediatric case-control cohort (Grove et al., 2019; NCBI GEO, 2020; ProteomeXchange Consortium, 2022; ProteomeXchange Consortium, 2025).

Facilities, funding, ethics.

This is a secondary analysis of de-identified public datasets, no wet-lab work is proposed. Computational work will be performed on a standard workstation plus cloud compute when necessary (budgeted). Public de-identified datasets are used, ethical approval is typically not required for such re-analyses. Any optional controlled-access genomics will be handled strictly under data-use agreements, and only permitted derived artifacts will be shared.

Estimated timeline. 6 months, preregistration-first, ending with an open pipeline release and a preprint-ready report.

Open science and data-sharing commitment.

We will preregister the protocol on ResearchHub and release code, containers, manifests, QC logs, and legally shareable derived tables (ResearchHub Foundation, 2026). Raw/individual-level restricted genomics will not be redistributed.

RFP fit. This proposal is aligned with the ResearchHub Foundation “Request for Proposals - Genomics and Proteomics”.

2. Introduction

Background.

Multi-omics technologies have produced large ASD-relevant datasets, but many are analyzed within narrow scopes. Modern, reproducible re-analysis and integrative methods enable extraction of new biological insights from existing data-particularly at the pathway/module level, which is more robust to platform differences than individual feature overlap.

Key literature and rationale.

Genomics: large ASD GWAS meta-analyses provide reproducible common-variant signal and enable summary-statistics-based gene/pathway mapping (Grove et al., 2019).
Microbiome: public pediatric 16S case-control cohorts exist with adequate sample size for external validation (GSE113690) (NCBI GEO, 2020).
Proteomics: pediatric ASD proteomics datasets hosted in ProteomeXchange / ProteomeCentral can be re-analyzed to derive differential protein signatures and pathway enrichment, serving as cross-omic anchors (ProteomeXchange Consortium, 2022; ProteomeXchange Consortium, 2025).

Research questions.

Can re-analysis of public pediatric proteomics identify ASD-associated protein/pathway modules that are internally consistent and interpretable (ProteomeXchange Consortium, 2022; ProteomeXchange Consortium, 2025)?
Do proteomics-enriched pathways converge with GWAS-derived genetic priors at the pathway/module level (Grove et al., 2019)?
Do microbiome signatures (taxonomic, optionally predicted function if sequences permit) show convergence with the same modules (NCBI GEO, 2020)?

Hypotheses.

H1: Proteomics re-analysis will yield ASD-associated pathways/modules that are statistically supported within each dataset (multiple-testing controlled) and biologically interpretable (ProteomeXchange Consortium, 2022; ProteomeXchange Consortium, 2025).
H2: Proteomics-enriched pathways will significantly overlap with GWAS-derived pathway priors beyond null expectation (Grove et al., 2019).
H3: Microbiome signatures will show pathway/module-level alignment with the same convergent modules, supporting cross-omic coherence even when individual-level genotype is unavailable (NCBI GEO, 2020).

Impact statement. Even if convergence is weak, this project will deliver a preregistered, open, reusable re-analysis pipeline and standardized dataset manifests/QC outputs that reduce friction for follow-up funding, publications, and translational studies.

3. Methods

3.1 Participants / Sample

Design. Secondary re-analysis of existing, de-identified public datasets, no recruitment of new participants.

Primary microbiome cohort (validation-ready).

GSE113690: fecal microbiome 16S V4 cohort; healthy controls (n=143) and ASD (n=143), total n=286; sequencing details provided in the GEO record (NCBI GEO, 2020).

Proteomics cohorts (public re-analysis, anchors).

PXD035099: plasma exosome proteomics dataset (ProteomeXchange Consortium, 2022).
PXD063396: dried blood spot proteomics dataset (ProteomeXchange Consortium, 2025).

Genomics priors (summary-statistics-based).

ASD GWAS summary statistics used for gene mapping and pathway priors (e.g., Grove et al., 2019).
Important constraint: individual-level PRS integration will be performed only if an appropriate genotype+phenotype dataset is legally accessible and consistent with open-sharing obligations; otherwise, genomics remains pathway priors only (no per-subject PRS).

Inclusion criteria (dataset-level).

Case/control label available.
Passes dataset-level QC.
Microbiome: OTU/taxon table available; proteomics: quantitative protein abundance available (raw or processed).

Exclusion criteria (preregistered, dataset-dependent).

Missing labels, failed QC.
If antibiotic metadata exists: exclude recent antibiotic exposure; if not available: include all and run sensitivity analyses, reporting limitations (NCBI GEO, 2020).

Sample size justification.

Microbiome n=286 enables stable train/test splitting and robust estimation of out-of-sample discrimination with confidence intervals (NCBI GEO, 2020).
Proteomics sample sizes are fixed by dataset; emphasis is on effect sizes, confidence intervals, and cross-dataset replication (ProteomeXchange Consortium, 2022; ProteomeXchange Consortium, 2025).

3.2 Materials and Procedures

Microbiome (GSE113690; 16S)

Start from public OTU/abundance tables and associated metadata (NCBI GEO, 2020).
Preprocessing: filtering low-prevalence features; normalization/compositional handling sensitivity analyses (relative abundance vs CLR with pseudocount); strict split hygiene to avoid leakage.
Functional inference:
- IF representative sequences/ASVs are available, THEN perform predicted functional profiling and report explicitly as “predicted function”;
- ELSE restrict to taxonomy-only models and treat functional inference as not feasible for this dataset (NCBI GEO, 2020).

Proteomics (PXD035099; PXD063396)

Use repository-provided quantification tables when available for fast reproducible re-analysis (ProteomeXchange Consortium, 2022; ProteomeXchange Consortium, 2025).
IF raw files are accessible and processing is feasible within budget/time, THEN run an open re-quantification workflow; ELSE analyze provided quantifications only.
Differential analysis + pathway enrichment (GO/Reactome) with FDR control.

Genomics priors (GWAS summary statistics)

Map GWAS summary statistics to genes and pathways (gene-set enrichment / pathway priors), producing a ranked “genetic priors list” for convergence testing (Grove et al., 2019).

3.3 Planned Analyses

Primary endpoint (POC success criterion). Evidence of cross-omic pathway/module convergence, defined as:

≥1 pathway/module significantly enriched in proteomics (FDR-controlled), and
significant overlap with GWAS-derived pathway priors beyond null expectation using permutation-based overlap tests controlling for gene-set size and pathway library structure (Grove et al., 2019; ProteomeXchange Consortium, 2022; ProteomeXchange Consortium, 2025).

Secondary endpoints. Microbiome baseline predictive performance on GSE113690: ROC-AUC, PR-AUC, calibration with confidence intervals; ablations for normalization choices (NCBI GEO, 2020).

Microbiome modeling plan (preregistered). Nested CV on training set + locked test evaluation; preregistered baselines (regularized logistic regression; tree-based baseline) with uncertainty via bootstrap/DeLong.

Proteomics analysis plan. Differential proteins → pathway enrichment (GO/Reactome), FDR control, pathway/module concordance across proteomics datasets where feasible (ProteomeXchange Consortium, 2022; ProteomeXchange Consortium, 2025).

Convergence testing plan. Define gene universes per assay; overlap statistics using hypergeometric + permutation null; report enrichment effect sizes and FDR.

Contingent IF–THEN decisions (preregistered).

IF batch effects detected, THEN include covariates where possible and run sensitivity analyses.
IF proteomics provides only processed tables, THEN focus on pathway-level robustness and transparently report limitations.
IF genotype access is incompatible with open outputs, THEN omit PRS and use GWAS priors only.
IF class imbalance >20%, THEN use stratified splits, class-weighted loss, and report PR-AUC.

3.4 Ethics and Data Management

Public de-identified datasets only, no attempt at re-identification. Preregistration on ResearchHub before confirmatory analyses, release code/containers/manifests/QC logs and legally shareable derived artifacts (ResearchHub Foundation, 2026).

4. Pilot Data

We have implemented a pilot microbiome ML pipeline (feature filtering + scaling + tree-based models). In a small pilot abundance dataset (n≈60; holdout split), ROC-AUC was high (≈0.9+). This pilot was not preregistered and likely optimistic due to small-sample variance, therefore, it is used strictly to demonstrate feasibility. Main confirmatory results will be preregistered and validated on the larger independent cohort GSE113690 (n=286) (NCBI GEO, 2020).

5. Budget

Cloud compute + storage: $2,500
Data engineering / harmonization: $1,000
Reproducibility packaging: $500
Dissemination: $1000

6. References

Grove, J., Ripke, S., Als, T.D., Mattheisen, M., Walters, R.K., Won, H., Pallesen, J., Agerbo, E., Andreassen, O.A., Anney, R., Awashti, S., Belliveau, R., Bettella, F., Buxbaum, J.D., Bybjerg-Grauholm, J., Bækvad-Hansen, M., Cerrato, F., Chambert, K., Christensen, J.H., et al. (2019) ‘Identification of common genetic risk variants for autism spectrum disorder’, Nature Genetics, 51(3), pp. 431–444. https://doi.org/10.1038/s41588-019-0344-8

NCBI GEO (2020) GSE113690: Gut microbiota analysis associated with autism spectrum disorders [16S rRNA]. Available at: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE113690

ProteomeXchange Consortium (2022) PXD035099: Proteomic analysis of plasma exosomes from children with autism spectrum disorder. Available at: https://proteomecentral.proteomexchange.org/cgi/GetDataset?ID=PXD035099

ProteomeXchange Consortium (2025) PXD063396: Comparative proteome analysis of the dried blood spot from children with autism spectrum disorder. Available at: https://proteomecentral.proteomexchange.org/cgi/GetDataset?ID=PXD063396

ResearchHub Foundation (2026) Request for Proposals – Genomics and Proteomics. Available at: https://www.researchhub.com/grant/5057/request-for-proposals-genomics-and-proteomics

Shaw, K.A., Bakian, A.V., Bilder, D.A., Durkin, M.S., Esler, A., Furnier, S.M., Hughes, M.M., Patrick, M., Pierce, K., Salinas, A., Shenouda, J., White, T., Zahorodny, W., and Maenner, M.J. (2025) ‘Prevalence and Early Identification of Autism Spectrum Disorder Among Children Aged 4 and 8 Years — Autism and Developmental Disabilities Monitoring Network, 16 Sites, United States, 2022’, MMWR Surveillance Summaries, 74(2), pp. 1–22. https://doi.org/10.15585/mmwr.ss7402a1

Open analysis and pathway-level multi-omics integration for ASD: GWAS-derived genetic priors, pediatric proteomics anchors, and gut microbiome signatures.

Open analysis and pathway-level multi-omics integration for ASD: GWAS-derived genetic priors, pediatric proteomics anchors, and gut microbiome signatures.

Funding Progress

Peer Reviews

Related Funding

Request for Proposals - Genomics and Proteomics

Nonprofit Recipient

Topics

DOI