Phenotype files should be in PLINK format. The first two columns should provide sample IDs, with subsequent columns providing values for each phenotype. The files can contain a header row, but then the first two elements must be named either FID & IID or ID1 & ID2.
When a phenotype file contains more than one phenotype (i.e., more than three columns), you can use --mpheno <integer> or --pheno-name <string> to specify which phenotype to analyse (the latter requires that the phenotype file has a header row). For example, --mpheno 2 tells LDAK to use the second phenotype, which is stored in the fourth column of the phenotype file, while --pheno-name P1 tells LDAK to use the phenotype labelled P1 (note that ). Alternatively, some functions (most notably LDAK-KVIK and REML) allow for testing of all phenotypes by adding --mpheno ALL.
Missing phenotypic values should be denoted by NA (note that while PLINK also treats -9 as missing, this is not the case in LDAK). Binary phenotypes should only take values 0 (control), 1 (case) or NA (missing). In general, LDAK excludes samples with missing phenotypes (the exception is when analysing multiple phenotypes, in which case LDAK replaces missing values with the mean value of the corresponding phenotype).
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Covariate files should also be in PLINK format (i.e., the first two columns provide sample IDs, with subsequent columns providing values for each covariate). If you have quantitative covariates, you should use --covar, while if you have categorical covariates, you should use --factors. Note that if a categorical covariate has U unique values, LDAK will (internally) replace it with U-1 indicator variables (LDAK will give an error if the total number of indicator variables is greater than half the sample size).
If providing quantitative covariates, you can use --covar-numbers <string> to specify a subset of covariates (using commas and/or dashes to specify multiple covariates). For example, --covar-numbers 1,2,4-6,8 tells LDAK to retain Covariates 1, 2, 4, 5, 6 and 8. If the covariate file has a header row, you can instead use --covar-names <string> (use commas to specify multiple covariates). For example, --covar-names Sex,Age tells LDAK to retain the covariates labelled "Age" and "Sex".
Missing covariate values should be denoted by NA (missing values are replaced by the mean value of the corresponding covariate).
