Genetically Adjusted Propensity Score Matching: A Comparison to Discordant MZ Twin Models

Ian A. Silver, Hexuan Liu, Joseph L. Nedelec

Research output: Contribution to journalArticlepeer-review


Discordant monozygotic (MZ) twin methodologies are considered one of the foremost statistical approaches for estimating the influence of environmental factors on phenotypic variance. Limitations associated with the discordant MZ twin approach generates an inability to estimate particular relationships and adjust estimates for the confounding influence of gene-nonshared environment interactions. Recent advancements in molecular genetics, however, can provide the opportunity to address these limitations. The current study reviews an alternative technique, genetically adjusted propensity scores (GAPS) matching, that integrates observed genetic and environmental information to adjust for the confounding of these factors in nonkin individuals. Simulations and a real data example were used to compare the GAPS matching approach to the discordant MZ twin method. Although the results of the simulated comparisons demonstrated that the discordant MZ twin approach remains the more robust statistical technique to adjust for shared environmental and genetic factors, GAPS matching - under certain conditions - could represent a viable alternative when MZ twin samples are unavailable. Overall, the findings suggest that GAPS matching can potentially provide an alternative to the discordant MZ twin approach when limited variation exists between identical twin pairs. Moreover, the ability to adjust for gene-nonshared environment interactions represents a potential advancement associated with the GAPS approach. The limitations of the approach, as well as polygenic risk scores, are also discussed.

Original languageEnglish (US)
Pages (from-to)24-39
Number of pages16
JournalTwin Research and Human Genetics
Issue number1
StatePublished - Feb 4 2022

All Science Journal Classification (ASJC) codes

  • Pediatrics, Perinatology, and Child Health
  • Obstetrics and Gynecology
  • Genetics(clinical)


Dive into the research topics of 'Genetically Adjusted Propensity Score Matching: A Comparison to Discordant MZ Twin Models'. Together they form a unique fingerprint.

Cite this