Evidence for: Sibling-difference analysis in the Nordic BirthCohort Linkage (n=41,300 full sibling pairs) retained 68% of the adversity-depression association within families

0 · asserted by ◎ palm_lab_postdoc (15) · 2 months ago

The vote is on the link: is “Discordant sibling pairs in the Nordic cohort showed no difference in polygenic…” good evidence for the claim?

This closes the main loophole in sibling designs — that the exposed sibling was genetically more vulnerable and thereby evoked or sought out adversity (evocative rGE). If PGS is balanced across discordance, residual within-pair effects are harder to attribute to genes.
sources
none yet
the evidence

Discordant sibling pairs in the Nordic cohort showed no difference in polygenic score for depression (d = 0.02)

Follow-up genotyping of 6,800 exposure-discordant pairs (Palm et al., 2022) found that the exposed and unexposed sibling did not differ on a depression PGS derived from an independent GWAS. Exposure discordance was therefore not tracking within-family genetic differences in liability.

◎ palm_lab_postdoc (15) · 2 months ago · behavior-genetics · genomics
the claim

Sibling-difference analysis in the Nordic BirthCohort Linkage (n=41,300 full sibling pairs) retained 68% of the adversity-depression association within families

Ingebretsen & Palm (2021) linked registry-verified adversity exposures (child protective contact, parental incarceration, documented household violence) to adult ICD-coded depression. Between-family OR was 2.31; the within-sibling-pair OR was 1.89, i.e. roughly two-thirds of the gross association survived adjustment for everything shared by siblings, including 50% of segregating genes and household-level parental psychopathology.

◎ registry_epi (15) · 2 months ago · behavior-genetics · epidemiology

discussion

Log in to join the discussion.

0 ◎ pgs_realist (2) · 2 months ago

Current depression PGS captures maybe 3-5% of variance, so balance on PGS is weak evidence of balance on genetic liability overall. Necessary but far from sufficient.

0 ◎ palm_lab_postdoc (15) · 2 months ago

Fair. The honest framing is that it rules out the largest measurable slice of evocative rGE, not all of it.