Merging bed bim fam files from chromosome 1 to 22 at one time

Question

[khn]

Hello,

I am trying to merge multiple bed bim fam files (each chromosome has separate file) into one file.

plink --allow-no-sex --bfile XX.chr1 --merge-list mylist.txt --make-bed --out XX.ch1-22

This does not work.

mylist.txt contains:

XX.chr2.bed XX.chr2.bim XX.chr2.fam
XX.chr3.bed XX.chr3.bim XX.chr3.fam
XX.chr4.bed XX.chr4.bim XX.chr4.fam
...
XX.chr22.bed XX.chr22.bim XX.chr22.fam

Thank you in advance!

[Devon Ryan]

Is plink giving you an error? Otherwise, please define "does not work".

[khn]

Yes, it says that "cannot open the file XX.chr2.fam". What would you think is wrong?

[Devon Ryan]

Does that file exist?

[khn]

Yes. I changed the txt file to csv file, but it still says "Error: Failed to open XX.chr2.fam."

[khn]

Message is as below... What would you think is wrong??

*** MB RAM detected; reserving *** MB for main workspace.
.ped scan complete (for binary autoconversion).
Performing single-pass .bed write (*** variants, *** people).
--file:
XX.ch1-22-temporary.bed
+
XX.ch1-22-temporary.bim
+
xx.ch1-22-temporary.fam
written.
Error: Failed to open XX.ch2.fam.

[bk11]

@khn How did you solve this problem? I am also new to plink. I am facing the same problem. How can we merge multiple binary files of different chromosomes using plink?

[T.F]

I have a similar trouble, but it seems to be different.

After conduct of command described in this page,

XX.ch1-22-merge.fam
XX.ch1-22-merge.missnp
XX.ch1-22.log

were generated, and no XX.ch1-22-merge.bed XX.ch1-22-merge.bim.

Is there any solutions?

Answer 1

[Shicheng Guo]

Here is the solution:

Be careful, before you merge different plink files, check the duplicate SNPs and remove them! and You'd better remove non-ATCG variants before merging, such as . in the dataset. Here, suppose you have chr1, chr2 .... and want to merge them.

chr1.bed
chr1.bim
chr1.fam
chr2.bed
chr2.bim
chr2.fam
....


rm mergelist.txt

for i in {1..22}
do
echo chr$i >> mergelist.txt
done

plink --merge-list mergelist.txt --make-bed --out G1000plink

[Ram]

Why the rm? OP's commands do not show an existing file by that name.

[Shicheng Guo]

See https://www.cog-genomics.org/plink/1.9/data#merge3 .

[Ram]

What is in there that justifies your rm command?

[h2obuffalo97]

Hi Ram, I believe the "rm mergelist.txt" is there so that it removes the mergelist.txt INCASE that file is already in existence. This code requires you to not have a mergelist.txt OR to have an empty mergelist.txt. So by "rm mergelist.txt", you make sure that the file doesn't exist prior to running your code.

[Ram]

If mergelist.txt does not exist, depending on other settings, there might be anything between a warning and script failure. Either test and remove, or use something like

seq 1 22 | xargs -I _num echo chr_num >mergelist.txt

so it overwrites any existing content. parallel > xargs > for and scripting an non-interactive careless rm can be dangerous.

Answer 2

[chrchang523]

See https://www.cog-genomics.org/plink/1.9/data#merge3 .

Answer 3

[mike313177]

Hi, it seems I just solved my problem on this issue. It may worth to check the file path in your merge-list text file, which means you need to make sure each path is really directed to the corresponding bed/bim/fam file, not just the file name, but the completed path to it. Hope it helps. Xiao