Including Genetic PCA Components in SEM

[ahmad valikhani]

Hi, 

I have a question regarding how to include principal components (PCs) derived from genetic population stratification in a structural equation model. Should these PCs be treated like typical covariates, correlated with the exogenous variables (~~)  while predicting the endogenous variables or should they be incorporated in a way that they predict both the exogenous and endogenous variables?

Thank you in advance for your time and help.

Best wishes,
Ahmad

[Elliot Tucker-Drob]

They should be included in the original GWAS that produced the summary statistics that you are using.

--
You received this message because you are subscribed to the Google Groups "Genomic SEM Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to genomic-sem-us...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/genomic-sem-users/147078a2-f08a-4338-86c9-225dd9157854n%40googlegroups.com.

[ahmad valikhani]

Thanks for your response. I already controlled for them when calcluating PRS. Now I want to control for their effects in my SEM analysis. BTW, I am using traditional SEM. 

[Elliot Tucker-Drob]

This group is for the Genomic SEM analytic framework and software, as introduced in this paper:

Grotzinger, A. D., Rhemtulla, M., de Vlaming, R., Ritchie, S. J., Mallard, T. T.,  Hill, W. D, Ip, H. F., Marioni, R. E., McIntosh, A. M., Deary, I. J., Koellinger, P. D., Harden, K. P., Nivard, M. G., & Tucker-Drob, E. M. (2019). Genomic structural equation modelling provides insights into the multivariate genetic architecture of complex traits. Nature Human Behaviour, 3, 513-525. [Nivard & Tucker-Drob jointly directed this work] Paper Supplement(Text/Figures) Supplement(Tables) GenomicSEM software

You shouldn't create factors from PGIs unless they are derived from fully independent discovery samples (in which case see my bioRxiv paper below). If they are from overlapping samples, then shared estimation error across the original GWASs will induce correlations among the PGIs that do not reflect shared genetic etiology.

In PGI analysis, best practice is indeed to correct for PCs at both the discovery and the prediction stages.

Regular SEM questions have come up on this group from time to time, so you can search the history, but we try to keep this group focused on Genomic SEM.

Tucker-Drob, E. M. (2017). Measurement error correction of genome-wide polygenic scores in prediction samples. bioRχiv. Link (to preprint)

To view this discussion visit https://groups.google.com/d/msgid/genomic-sem-users/5f7e4f62-2b6d-4f34-b0ff-8da5a758b60bn%40googlegroups.com.