Research questions
Do genetically elevated maternal prenatal inflammation (CRP, IL-6) and insulin-related traits (BMI, fasting glucose) causally raise offspring risk of autism spectrum disorder (ASD) and atypical neuro-development?
Does foetal brain imaging (callosal area, arcuate-fasciculus FA) mediate >10 % of this effect?
Can a maternal-foetal polygenic risk score (PRS) predict childhood-diagnosed ASD in 1,000 UKB mother-child pairs and 40 k ALSPAC replicates?
Objectives
Build maternal immune-metabolic PRS (CRP, IL-6, TNF-!, BMI, HbA1c) and offspring ASD PRS (iPSYCH 2023 GWAS) in 1,017 genotyped UKB trios.
Quantify autistic-like neuro-anatomy on 50 k diffusion MRI: corpus-callosum area, superior-temporal-gyrus thickness, arcuate/uncinate FA; repeat in 7 T sub-cohort (N=1 k).
Estimate causal effects with two-sample Mendelian randomisation (MR) using maternal genotypes as instrumental variables; falsify with paternal negative-control MR.
Derive and validate a combined maternal-foetal PRS plus neonatal brain metric to predict ICD-10 F84 ASD at age !3 years; calibrate sensitivity/specificity in ALSPAC trios.
Release open Python/R package “UKB-ASD-MR” with Docker notebooks to accelerate prenatal ASD prevention trials.
Scientific rationale
Maternal infection/GDM doubles ASD risk, yet causality and mediators remain unproven. UKB is the only biobank offering (i) 1 k verified mother-child trios, (ii) 50 k multimodal MRI, (iii) linked HES ASD diagnoses, and (iv) existing GWAS for instrument selection. Demonstrating causal immune-insulin pathways would identify modifiable prenatal targets and yield an imaging-biomarker risk tool, meeting Nature Psychiatry / Molecular Psychiatry priorities and UKB’s translational mission.
