I've got some plink files (.bed .bim .fam) that allegedly contain genetic variation data for many (human) individuals. I would like to analyse this data using my own tools but can't see how to parse the plink format. Having never used plink I can't easily see how to get it to simply dump the data to an easily parseable file, e.g. in VCF format. But any reasonable well-documented easily-readable format will do.
I would really appreciate any tips from plink users out there who no doubt know the best way to go about it and could save me hours messing about with it. plink also seems to do some kind of filtering on variant frequency (removing rare variants?) by default which I don't want. I just would like to read the data.
Use plink/seq (http://atgu.mgh.harvard.edu/plinkseq/)
Note: you need the reference allele file to have plink to specify reference allele
#!/bin/sh
# 1. have plink binary to specify reference allele
plink --noweb --bfile $plink_file --reference-allele $ref_Allele_file --make-bed --out $plink_file_modified
# 2. create plinkseq project
pseq $pseq_project new-project
#3. load plink file into plink/seq
pseq $pseq_project load-plink --file $plink_file_modified --id $plink_file_modified
#4. write out vcf file
pseq $pseq_project write-vcf | gzip > $plink_file_modified.vcf.gz
hope it helps!
I think this could help you, PGDSpider, An automated data conversion tool for connecting population genetics and genomics programs
http://www.cmpg.unibe.ch/software/PGDSpider/
This may help you:
http://bhoom.wordpress.com/2012/04/06/convert-plink-format-to-vcf/
Use plink/seq (http://atgu.mgh.harvard.edu/plinkseq/)
Note: you need the reference allele file to have plink to specify reference allele
#!/bin/sh
# 1. have plink binary to specify reference allele
plink --noweb --bfile $plink_file --reference-allele $ref_Allele_file --make-bed --out $plink_file_modified
# 2. create plinkseq project
pseq $pseq_project new-project
#3. load plink file into plink/seq
pseq $pseq_project load-plink --file $plink_file_modified --id $plink_file_modified
#4. write out vcf file
pseq $pseq_project write-vcf | gzip > $plink_file_modified.vcf.gz
hope it helps!
Thanks guys this is exactly what I needed. Your detailed explanation is particularly appreciated Mehdi.
I was already knocking on the door of using plink/seq but was looking at some older documentation and couldn't find where to get it from. I also downloaded vcftools but was getting bogged down in too many options. PGDspider looks worth a try too.
Thanks again.
Alternatively, you can simply export the data from plink in what's known as LINKAGE format files (simple text files):
plink --bfile $plink_file --recode --out $plink_file_modified --noweb
This will create two files (${plink_file_modified}.ped and ${plink_file_modified}.map) - the .ped file contains information for each individual in the original files, in the following format:
column 1: family ID
column 2: individual ID
column 3: father ID
column 4: mother ID
column 5: sex
column 6: phenotype (or case/control status)
column 7+: genotypes for all SNPs
(essentially replicating the .fam file in the original format, but adding the genotypes for each person to the appropriate line)
The .map file contains information on the markers (SNPs) in the following format:
column 1: chromosome
column 2: marker ID
column 3: genetic location (in cM)
column 4: physical location (in bp)
(essentially replicating the .bim file from the original format)
This format is arguably even easier to parse than VCF format (and there are numerous tools out there to parse LINKAGE-format data - from the long history of linkage and population-based genetics studies)...
Hope that helps!
The free Mega2 (http://watson.hgen.pitt.edu/docs/mega2_html/mega2.html) software will convert PLINK input files to a wide variety of other data formats. Of these, the Mega2 annotated format is easily readable in R (if you don't have too many markers).
I arrived a bit late, but if it helps anybody, plink can now output vcf directly
http://genome.sph.umich.edu/wiki/Minimac3_Cookbook_:_Converting_Files_to_VCF
It contains a list of SNP IDs and the allele to be set as A1, e.g.
rs10001 A
rs10002 T
rs10003 T
...
# for dbSNP_BUILD_ID=142 (reference=GRCh37.p13):
wget ftp://ftp.ncbi.nih.gov/snp/organisms/archive/human_9606_b142_GRCh37p13/VCF/All.vcf.gz
zcat All.vcf.gz | grep -v "^#" | cut -f 3,4 > reference_allele_GRCh37.txt
# for dbSNP_BUILD_ID=144 (reference=GRCh38.p2):
wget ftp://ftp.ncbi.nlm.nih.gov/snp/organisms/human_9606/VCF/All_20150603.vcf.gz
zcat All_20150603.vcf.gz | grep -v "^#" | cut -f 3,4 > reference_allele_GRCh38.txt
Hi
I am going to creat marker LD plot (SNP and DArT) marker)
How I can do it in R?
Please let me know for input file format and procedure.
Thanks,
Reza
Hello everyone,
I have received the imputed data in the form of plink format (.bed, .bim, and .fam) for ch 1 to 22 separately.
This is my first time to work on imputed data distributed in multiple chromosomes. Please let me know how I can combine these files to perform further association analysis.
Thanks in advance
M
Hi Mahantesh,
You can do the following:
(echo -en "\x6C\x1B\x01";
for chr in {1..22};
do tail -c +4 imputed_chr"$chr".bed;
done) > imputed_ALL.bed
You can proceed with the .bim files the same way by doing
for chr in {1..22};
do
cat imputed_chr"$chr".bim;
done > imputed_ALL.bim
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics