SmartPCA setup

1. Created a smartpca reference population list

File used:

smartpca_refs.txt

This file contained the modern reference populations used to define the PCA axes, along with the modern target sample included as a reference in this run.

2. Created a convertf keep list

File used:

convertf_keep_projected.txt

This file contained:

all populations included in the smartpca reference list
additional ancient samples retained in the subset for projection

The ancient samples were kept in the dataset but not included in the smartpca reference list.

3. Prepared a convertf parameter file

File:

convertf_subset.par

Structure:

genotypename:    merged_input.geno
snpname:         merged_input.snp
indivname:       merged_input.ind
outputformat:    EIGENSTRAT
genotypeoutname: pca_subset.geno
snpoutname:      pca_subset.snp
indivoutname:    pca_subset.ind
poplistname:     convertf_keep_projected.txt
hashcheck:       NO

4. Ran convertf

convertf -p convertf_subset.par > convertf_subset.log 2>&1

This generated the PCA subset used in the next step.

5. Prepared a smartpca parameter file

File:

smartpca.par

Structure:

genotypename: pca_subset.geno
snpname:      pca_subset.snp
indivname:    pca_subset.ind

evecoutname:  pca_output.evec
evaloutname:  pca_output.eval

poplistname:  smartpca_refs.txt

lsqproject:   YES
numoutevec:   10
numthreads:   10
numoutlieriter: 0

6. Ran smartpca

smartpca -p smartpca.par > smartpca.log 2>&1

7. Output files generated

pca_output.evec
pca_output.eval

This run used the populations in smartpca_refs.txt to define the PCA axes, while samples retained through convertf_keep_projected.txt but absent from the smartpca reference list were projected.