CCA(gwas)

This is a short review of a multivariate test of associations[@ferreira2009multivariate] implemented in PLINK[@Purcell2007559].

Remained Questions:

1. Derivation of the statistics still don’t understand;

---

Canonical Correlation Analysis(CCA) [@wiki_cca]

Summary

• CCA is the multivariate generalization of the Pearson product-moment correlation;
• CCA finds linear combinations of the two sets of random variables which have maximum correlation with each other;

Notation and data

• Two sets of random variables $X =(x_1,...,x_n)',Y =(y_1,...,y_m)'$;
• data would be two matrix with column be sample size $p$;
• Problems of correlations: $$Y \leftrightarrow X$$
• cross covariance: ∑XY