This is a test version of Biostars. For the public version, visit https://www.biostars.org.
use of 2 covariate files for association analysis in plink

Dear all, I want to use 2 covariate file in my association analysis in plink, is this possible ?, (without combining all the covariates in 1 file) Is it valid to use ? --bfile file \ --logistic \ --covar /covarfile1.txt \ --covar-number 1-3 \ --covar/covarfile2.txt --covar-number2-3 \ --out filename Thanks

cov plink association
• 4,482 views • link
updated by 1311703846 • 1311703846 • written by SKS • SKS

Please use:
-A more informative title: adding stuff like "association analysis" and "plink" would be good
-More precise tags: for example "plink"

• 0 views • link
written by WouterDeCoster WouterDeCoster

1 answer

No; you'll need to merge the files with a short script.

• 0 views • link
written by chrchang523 chrchang523

Many thanks for your response. Could you please advice how the covariate files can be merged based on FID and IID in bash/ plink? (The individuals can be uniquely identified only by using both FID and IID information). And not al individuals in first covariate file are present in the second covariate file and vice versa.

• 0 views • link
written by SKS • SKS

With a bash script:

A. Merge the FID and IID fields with something like

cat orig_covar_1.txt | awk '{out=$1"_"$2; for(i=3;i<=NF;++i){out=out"\t"$i}; print out}' > covar_merged_ids_1.txt

cat orig_covar_2.txt | awk '{out=$1"_"$2; for(i=3;i<=NF;++i){out=out"\t"$i}; print out}' > covar_merged_ids_2.txt

If underscores are already present in your file, replace "_" with some other character that isn't used.

B. Use the Unix "join" command (https://en.wikipedia.org/wiki/Join_(Unix) ) to merge covar_merged_ids_1.txt and _2.txt.

C. Split the merged IDs back into FID/IID with e.g.

cat covar_joined_merged_ids.txt | tr '_' '\t' > covar_joined.txt

• 0 views • link
written by chrchang523 chrchang523

Thank you very much.

• 0 views • link
written by SKS • SKS

Suppose your files doesn't has header row, and using space as separator. then the following code could be applied for your question.

awk 'FNR==NR{j=$1 FS $2;arr[j]=$3;for(i=4;i<=NF;i++){arr[j]=arr[j] FS $i};next}{m=$1 FS $2;print $0,arr[m];delete arr[m]}END{for(k in arr){print k,arr[k]}}' test1 test

where test1 looks like

a b c q
a c q x
a i o a
b c q q
b d p e
b w a h

test looks like

a b c e
a c e z
b c a w
b w a w

and the result would be like

a b c e c q
a c e z q x
b c a w q q
b w a w a h
a i o a
b d p e

the above awk code match first two columns in test1 and test, and then merge two files together. and there are two rows in test1 that do now appear in test ,so i print these lines in the end of the output.

• 0 views • link
written by 1311703846 • 1311703846

Log in to answer this question.

More posts like this