Identification of twin pairs from large population-based samples

Area of Science:

• Twin studies
• Population genetics
• Educational research

Background:

• Twin studies are crucial for understanding genetic and environmental influences.
• Ascertainment of twins, often via registries and zygosity determination, is the first step.
• Rising twin birth rates necessitate efficient identification methods in large datasets.

Purpose of the Study:

• To quantify the number of twin pairs detectable in a large, population-based longitudinal survey.
• To assess the feasibility and accuracy of twin identification using available identifiers.
• To evaluate the potential of these identified twins for future research.

Main Methods:

• Utilized a longitudinal survey dataset of 284,945 primary school pupils in the Netherlands.
• Employed a stringent set of identifiers including coded name, date of birth, school, grade, and survey year.
• Applied intraclass correlation of phenotypes as a quality control for twin identification.

Main Results:

• Detected 2865 twin pairs, representing 2.01% of the sample, using strict criteria.
• Showed that relaxing criteria increased false positives.
• Estimated detection of over 80% of true twin pairs in the sample with stringent criteria.

Conclusions:

• Identification of twin pairs from large population samples is feasible, rapid, and accurate with appropriate identifiers.
• Twins from population-based samples are a valuable resource for research on twin vs. non-twin differences and genetic/environmental factors.
• This approach enhances the utility of existing large datasets for twin research.