Sample Dataset

504 EAS individuals from 1000 Genomes Project Phase 3 version 5

• CHB: Han Chinese in Beijing, China
• JPT: Japanese in Tokyo, Japan
• CHS: Southern Han Chinese
• CDX: Chinese Dai in Xishuanagbanna, China
• KHV: Kinh in Ho Chi Minh City, Vietnam

Url: http://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/

Genome build: human_g1k_v37.fasta (hg19)

Genotype Data Processing

• Selected only autosomal variants
• Split multi-allelic variants
• Variants were normalized
• Remove duplicated variants
• Selected only SNP (ATCG)
• Selected 2% rare SNPs (plink --mac 2 --max-maf 0.01 --thin 0.02)
• Selected 15% common SNPs (plink --maf 0.01 --thin 0.15)
• Converted to plink bed format and merged to a single file
• Randomly added some missing data points

Download

You can also simply run download_sampledata.sh in 01_Dataset and the dataset will be downloaded and decompressed.

./download_sampledata.sh

or you can manually download it from this link.

Unzip the dataset unzip -j 1KG.EAS.auto.snp.norm.nodup.split.rare002.common015.missing.zip, and you will get the following files:

1KG.EAS.auto.snp.norm.nodup.split.rare002.common015.missing.bed
1KG.EAS.auto.snp.norm.nodup.split.rare002.common015.missing.bim
1KG.EAS.auto.snp.norm.nodup.split.rare002.common015.missing.fam

Phenotype Simulation

Phenotypes were simulated using GCTA with the 1KG EAS dataset. GCTA provides a function to simulate GWAS data based on observed genotype data.

Quantitative Trait Model

For a quantitative trait, phenotypes are simulated using the simple additive genetic model:

where: - \(y\) is the phenotype vector - \(W\) is the genotype matrix (coded as 0, 1, or 2 for the number of effect alleles) - \(u\) is the vector of SNP effects - \(\epsilon\) is the vector of residual effects

Simulation procedure:

1. Causal variant effects: Given a set of SNPs assigned as causal variants, the effects (\(u\)) are generated from a standard normal distribution: \(u \sim N(0, 1)\)

2. Residual effects: The residual effects (\(\epsilon\)) are generated from a normal distribution with mean 0 and variance \(\sigma^2_g(1/h^2 - 1)\): $\(\epsilon \sim N(0, \sigma^2_g(1/h^2 - 1))\)$

where: - \(\sigma^2_g\) is the empirical variance of \(Wu\) (genetic variance) - \(h^2\) is the user-specified heritability

Case-Control Study Model

For a case-control study, GCTA assumes a threshold-liability model:

1. Disease liabilities are simulated in the same way as phenotypes for a quantitative trait using the model \(y = Wu + \epsilon\)

2. Any individual with disease liability exceeding a certain threshold \(T\) is assigned as a case, and a control otherwise

3. The threshold \(T\) is determined from the normal distribution truncating the proportion \(K\) (disease prevalence)

Example Command

gcta  \
  --bfile 1KG.EAS.auto.snp.norm.nodup.split.rare002.common015 \
  --simu-cc 250 254  \
  --simu-causal-loci causal.snplist  \
  --simu-hsq 0.8  \
  --simu-k 0.5  \
  --simu-rep 1  \
  --out 1kgeas_binary

$ cat causal.snplist
2:55620927:G:A 3
8:97094292:C:T 3
20:42758834:T:C 3
7:134326056:G:T 3
1:167562605:G:A 3

References

• 1000 Genomes Project Consortium. (2015). A global reference for human genetic variation. Nature, 526(7571), 68.
• Yang, J., Lee, S. H., Goddard, M. E., & Visscher, P. M. (2011). GCTA: a tool for genome-wide complex trait analysis. The American Journal of Human Genetics, 88(1), 76-82.