Most reported genetic associations with general intelligence are probably false positives

Area of Science:

• Behavioral genetics
• Neurogenetics
• Psychometrics

Background:

• General intelligence (g) and behavioral traits exhibit heritability.
• Previous research suggested associations between g and specific single-nucleotide polymorphisms (SNPs) in candidate genes.
• Replication is crucial for validating genetic findings in complex traits.

Purpose of the Study:

• To replicate previously reported associations between general intelligence (g) and 12 specific genetic variants.
• To assess the reliability of candidate gene associations for g using large, independent datasets.
• To evaluate the current candidate gene approach in behavioral genetics.

Main Methods:

• Utilized three independent longitudinal study datasets (N=5,571, 1,759, 2,441).
• Tested associations between g and 12 candidate gene SNPs (e.g., DTNBP1, COMT, BDNF, APOE).
• Performed power analyses and confirmed known genetic associations for Alzheimer's disease and BMI as positive controls.

Main Results:

• Only 1 out of 32 tests yielded a nominally significant association between candidate genes and g.
• Power analyses indicated an expected 10-15 significant associations if effects were real.
• SNP-based relatedness confirmed that common genetic variation accounts for approximately 50% of the variance in g.

Conclusions:

• The candidate gene approach is insufficient for explaining the genetic architecture of general intelligence.
• Molecular genetics in psychology and social science necessitates broader genomic strategies beyond single candidate genes.
• Current findings highlight the need for novel methodologies in behavioral genetics research.