Table 1 Summary of the research question, data sources used, as well as key strengths and limitations of each methodological approach applied in the present study

Nationwide registry-based cohort study in Sweden

Are parental diagnoses of IBD associated with autism in the children?

Medical and administrative registers

Large, diverse total population, intergenerational sample; prospective recording of data; low rate of loss to follow up; large availability of confounder data

Unmeasured confounding; exposure misclassification

Linkage Disequilibrium score regression

Is there a shared genetic background between IBD and autism?

GWAS summary data

Use of GWAS summary data instead of twin data or individual-level data maximizes sample sizes and power; indicates genetic correlation due to linkage disequilibrium or pleiotropy

Cannot assess causality

PRS analysis in the ALSPAC cohort

Is maternal genetic liability for IBD associated with childhood broad autism phenotype?

GWAS summary data and individual-level genotype and phenotype data

Estimates the underlying genetic liability for IBD, regardless of diagnosis; allows the refinement of the exposure used in the context of an observational study (potentially overcoming exposure misclassification of an observational study); can indicate potentially genetically transmitted versus in-utero effects through the assessment of the maternal versus child’s underlying genetic liability for IBD; large birth cohort; prospectively collected information on the outcome phenotype

Cannot decipher whether the identified associations are causal or instead owing to pleiotropy; polygenic risk scores at lower P-value thresholds might not adequately capture the exposure phenotype; attrition can influence association estimates

Two-sample MR

Does genetic liability to IBD have a causal effect on autism?

GWAS summary data, exposure proxied by variants robustly associated with the exposure

Using common genetic variants as instruments for IBD allows the assessment of causal effects; allows the assessment of reverse causation; allows the assessment of the influence of pleiotropy

Cannot decipher whether the identified causal effect is of parental origin; can be biased by dynastic effects and assortative mating