Input files

To run the multivariate association test, you'll need the standard Plink PED/MAP or BED/BIM/FAM (binary) filesets as well as an alternate phenotype file with the phenotypes to be analysed. The formats for these files are described in detail in the Plink website, and examples that can be used to try the multivariate association test can be downloaded below.

Plink file type File name File size
PED file ex1.ped 43 KB
MAP file ex1.map 250 bytes
Alternate phenotypes file ex1.phen 34 KB

This fileset contains data for 500 unrelated individuals, genotyped at 18 SNPs and with phenotype information available for 10 quantitative traits. The PED file has dummy sex and affection status fields (ie. all = 1). The affection status field of the PED file is not used for the multivariate analysis. Instead, the multivariate test will analyse by default all phenotypes supplied in the alternate phenotype file. To analyse only a subset of these, use the --pheno-name or --pheno-number option (see usage).

The first five data lines of the alternate phenotypes file ex1.phen look like this:

       [contents of ex1.phen]
	FID IID T1 T2 T3 T4 T5 T6 T7 T8 T9 T10
	1 1 -0.333 -1.17 1.655 -0.232 -0.294 -0.654 -1.926 0.238 -0.171 -0.222
	2 1 0.812 0.941 0.345 -0.224 0.468 0.656 -0.079 0.209 0.202 1.163
	3 1 0.648 0.275 0.09 -0.313 -0.608 -1.296 1.133 -0.192 -0.459 1.088
	4 1 0.647 -0.249 0.02 -0.495 0.226 -0.87 1.482 -0.355 1.811 0.097
	5 1 1.14 -0.676 -0.001 0.258 1.806 -0.897 -0.195 -0.092 0.286 1.036

The file has one line per individual, starting with the family (FID) and individual (IID) identifiers, followed by a value for each of the traits measured, in this case 10 traits. A -9 is by default considered a missing value, but this behavior can be modified with the --missing-phenotype option. A header is not required but is usually helpful.