Imputed genetic data - genetics-of-dna-methylation-consortium/godmc_phase2 GitHub Wiki

Genetic data

Data pre-processing

Prior to running the GoDMC pipeline we require that genetic data must be

Imputed to HRC reference panel, ideally v1.1. The Haplotype Reference Consortium offers a free imputation service (including HRC) which you can use here or here. By far the easiest way to do imputation now is to use this service. The pipeline does work on 1000G as well but we have a strong preference for HRC imputation.
Variant positions in build 37
Filtered to have MAF > 0.01 and imputation quality score > 0.8
IMPORTANT: Converted to best guess binary plink format without a probability threshold
All remaining SNPs combined into a single fileset (i.e. not a separate fileset for each chromosome)
No spaces in the file names
All autosomes
Chromosome X is optional and should be coded as 23

Fam file and Sample IDs

The fam file should meet the standard plink structure and not contain a header.
The first column of the .fam file are the family IDs. If you have unrelated data then this can be the same as the second column (individual IDs). If you have related data, please ensure your family IDs correctly capture these relationships.
IMPORTANT: Please ensure that the second column of the .fam file (individual IDs) contains unique sample IDs. These sample IDs should be the same IDs that are used in the covariate and methylation data. Please also ensure that the individual IDs don't contain any underscores. Individual IDs cannot be 0.
Column 3 and 4 are the fatherID and the motherID. If you don't have parent data please set these IDs to 0.
Column 5 is sex; 1 is male, 2 is female.
Please note that you need to set the phenotype (column 6) to missing i.e. "-9".

Example fam file for unrelated data

1A 1A  0  0  2 -9
2A 2A  0  0  2 -9    
3A 3A  0  0  2 -9

Example fam file for related data (twins, families, mother-offspring etc.)

1 1A  0  0  2 -9
1 1B  0  0  2 -9    
2 2A  0  0  2 -9

Bim file and allele coding

The bim file should meet the standard plink structure and not contain a header.
Column 1 is chromosome. Chromosomes should be coded as 1-23, where chromosome X is coded as 23.
Column 2 is variant name. The pipeline also changes the SNV ids for you (second column) so you don't need to worry about marker names.
Column 3 should be 0.
Column 4 should be basepair position for build 37.
Column 5 and 6: IMPORTANT: ensure that your alleles coded using the original impute2 or minimac coding. To perform a meta-analysis across cohorts, alleles should be matching across cohorts.
Please note HRC has SNPs only and no INDELs. The pipeline will remove INDELs from those that have 1000G imputed data.

Bed file

IMPORTANT: We use best guess genotypes without a probability threshold in our pipeline. Please make sure you don't filter on a probability threshold as some software packages can't handle missingness properly and will set missing genotypes to the genotypic mean. We use best guess genotypes in our analyses. Please see below how you can prepare best guess files.

Imputation quality

We also require imputation quality scores for each SNP. Some instructions on how to get imputed data into the desired format including your imputation quality file are below.

    SNP MAF Info
    rs1 0.02 0.88