Ancestry PCA

Ancestry analysis and evaluate the similarity or difference between populations or superpopulations based on 1000 genomes data by using principal component analysis

method - PCA in plink tool to reduce dimension (10 components) and select 2 principal components that maximizes the variance explained to visualize results.

• Data: Chromosome 20 genotypes (vcf file) and mapping population (csv file). Default directory: database/

• Information: The 1kGP data provide 2504 samples which were collected with specific populations as follows:

  Superpopulation	Population	Details
  EAS	CHB	Han Chinese
  	JPT	Japanese
  	CHS	Southern Han Chinese
  	CDX	Dai Chinese
  	KHV	King Vietnamese
  	CHD	Denver Chinese
  ---	---	---
  EUR	CEU	CEPH
  	TSI	Tuscan
  	GBR	British
  	FIN	Finnish
  	IBS	Spanish
  ---	---	---
  AFR	YRI	Yoruba
  	LWK	Luhya
  	GWD	Gambian
  	MSL	Mende
  	ESN	Esan
  ---	---	---
  AMR	ASW	African-American SW
  	ACB	African-Caribbean
  	MXL	Mexican-American
  	PUR	Puerto Rican
  	CLM	Colombian
  	PEL	Peruvian
  ---	---	---
  SAS	GIH	Gujarati
  	PJL	Punjabi
  	BEB	Bengali
  	STU	Sri Lankan
  	ITU	Indian

• Output: 2D plot with first and second variance explained components. Default directory: output_figure/

1. Create conda environment with required packages:

conda env create -f ancestry.yml
conda activate ancestry

2. Run ancestry_pca.py:

# examples:

# with -s (superpopulation)
python ancestry_pca.py -s "EUR EAS AFR SAS AMR"
# with -p (population)
python ancestry_pca.py -p "ASW ACB KHV JPT STU CEU ESN"
# check percentile ancestry of samples, example: HG01595 HG01607 HG03702 HG03078
python ancestry_pca.py -s "EUR EAS AFR SAS AMR" -u "HG01595 HG01607 HG03702 HG03078"

# Output .png file in default directory 'output_figure'

• Flags:

    # Note: choose only population (-p) or superpopulation (-s)
    
    # choose population to plot
    -p --population
    # choose superpopulation to plot
    -s --superpopulation
    # check percentile ancestry of samples
    -u --user
    
    # view all flags
    python ancestry_pca.py --help