My raw data is simple: One column of SNP IDs followed by two columns for each locus where 1 is allel A and 2 is allel B. A homozogote would thus have 1 1 or 2 2 in the two first columns and a heterozygote 1 2 or 2 1. How can I convert this table into plink binary PED files for the analysis in GCTA? It seems like many of the variables (Family ID, Paternal ID, Maternal ID, Sex etc) are missing in my dataset and I also don´t know whether that could cause any problems.
You may have to construct the file by yourself using a script or so.
Be aware that Plink needs all these Family ID, Individual ID, Paternal ID, Maternal ID, Sex, Phenotype. You could use Family ID to put population names, individual ID for the individual names and the rest can be 0s. Column 7 onwards belong to the genotypes. That should work for the PED file.
You'll need also a MAP file. Use 0s for the Chr column, snp IDs for the second column and, if you don't have distances, use 0s for the third and a list of numbers (and be aware that you did that) for the fourth column.
http://pngu.mgh.harvard.edu/~purcell/plink/data.shtml
Hope it helps.
Hi Martin,
One way you can do this is by manipulating your data into PLINK PED format, then use PLINK to convert it into a binary PED.
To get you data into PLINK PED format I would use R. If none of your individuals are related, you can simply use the IID as FID for each individual. Then create columns for Mat ID, Pat ID and Sex filled with "0". I have attached an example with all phenotype values as "1" and two genotypes (example.ped). Remember there should not be column names.
You also need to create a .map file containing info on the SNPs. There are details of the .map format on the PLINK website. I have attached an example (example.map).
Once you have this, you can use PLINK to convert it to a binary PED file like this:
./plink --file example --make-bed --out example
I have never run GCTA myself but I don't think the missing values will matter as it calculates relatedness using SNP data. With out sex data you won't be able to control for sex in your analysis. If this doesn't matter for your analysis then it should be fine.
Hope this helps.
Oliver
thanks to everyone. These are really helpful answers. I just wonder about the fourth column in the MAP file (Base-pair position (bp units), I think its meant to be relative to the chromosome). Is it ok to just create a variable with ANY numbers in it?
It shouldn't be that difficult to annotate your rsIDs with base pair positions. Find out what array was used and download the annotation file. I think you will save yourself trouble in the future by getting this info right now. For one, GCTA - GREML needs linkage disequilibrium to be controlled for which may use the distance between SNPs.
Dear Sameera,
Do you have genetic data that you want to convert into PLINK format? If so what format is it in currently in? If you can send me an example of the data then I might be able to help.
Hey,
Use fcgene, I recommend! is the best way to convert plink database to all formats for genotyped database. Moreover is easy to use.
Oliver Pain,
I had txt format od GWAS data, then I edited it on R. Now I want to save it on PLINK format. Can I do that? Do you can share with me the scritp to do that ? Thank you @
Did you edit the ped or map file?
Can you share part of the file so that its easier to judge.
No, I had edit a TXT file in R. So I want to transform that to PLINK format...
Sorry, how can I share it with you here?
It depends on the file's original format. I wonder what is the source of the genotypes. What was its original format? Was it in VCF?
I ask because plink should require mapping information which is saved in the VCF format but I am not sure in your case. You can share part of the file using the add files icon below.
Hi,
There is a simple way to convert STRUCTURE format to PLINK?
thanks
Hi,
Try pdgspider. It converts between various formats.
http://www.cmpg.unibe.ch/software/PGDSpider/
Hi Mohammed,
Can PGDSpider convert in .bed, .bim and .fam set of files of PLINK?
thanks!!
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools