Mixed-Models

Mixed models with polygenic effects

The repository is motivated from a contribution to genetic analysis workshop 14 (GAW14).

Several aspects (in subdirectories) of the models are exposed,

  1. Mixed modeling with whole genome data (JPS) provides an overview and is derived from work on analysis of family data where relationship between relatives is commonly used and recently mirrored in population-based samples with a genomic relationship matrix (GRM) from whole genome genotypes.

  2. Bayesian linear mixed model with polygenic effects (JSS) adds Bayesian implementation using software such as OpenBUGS, JAGS and Stan plus specialised software.

  3. GxE interaction (GxE) is considered via both frequentist and Bayesian approaches above.

They have generic implications in assessment of SNP-trait association including risk prediction and Mendelian randomisation analysis, given that increasing number of variants are identified but with inter-popuation fluctuations. These variants are often used collectively as polygenic risk scores.

Exposition

This regards p values during analysis of BiSeq data, as it is shown here, https://github.com/jinghuazhao/QTR.

EnvWAS

This is associated with the CLARITE package, R package (https://github.com/HallLab/clarite), Python package (https://github.com/HallLab/clarite-python and https://halllab.github.io/clarite-python/) with index (https://pypi.org/).

References

Asar O, Bolin D, Diggle PJ, Wallin J (2018). Linear Mixed-effects models for non-Gaussian repeated measurement data. arXiv:1804.02592, https://arxiv.org/abs/1804.02592

Casale FP (2016). Multivariate linear mixed models for statistical genetics (Doctoral thesis). https://doi.org/10.17863/CAM.13422.

Gad AM, EL-Zayat, NI (2018). Fitting Multivariate Linear Mixed Model for Multiple Outcomes Longitudinal Data with Non-ignorable Dropout. International Journal of Probability and Statistics, 7(4): 97-105, DOI: 10.5923/j.ijps.20180704.01

Lucas AM, Palmiero NE, Orie D, Ritchie MD, Hall MA (2019). CLARITE facilitates the quality control and analysis process for EWAS of metabolic-related traits . Frontiers in Genetics 10:01240. doi: 10.3389/fgene.2019.01240

Pain O, GenoPred, https://opain.github.io/GenoPred/index.html (Diverse Ancestry)

Zhao JH (2005). Mixed-effects Cox models of alcohol dependence in extended families, BMC Genetics 6 (Suppl 1):S127.

Zhao JH, Luan JA (2012). Mixed modeling with whole genome data. Journal of Probability and Statistics, Volume 2012, 1-16.

Zhao JH, Luan JA, Congdon P (2018). Bayesian linear mixed models with polygenic effects. Journal of Statistical Software 85(6).

Zhao JH, Scott R, Luan JA, Sharp S, Langenberg C, Wareham NJ (2018). Polygenic and interaction effects in Type-2 diabetes – The InterAct study (unpublished manuscript).