Fastphase To Plink Format

Question

[kumar.vinod81]

I've imputed my plink.ped file using fastPHASE. It was imputed successfully. But know I want to convert this fastPHASE file back into PLINK format. Any idea...

fastPHASE format (2 individuals and 11 SNPs, row wise mutation)

ID 5906291001R01C01 1 1 1 0 1 1 1 1 1 0 1 1 1 1 0 1 1 1 1 1 0 1 
ID 5906291001R01C02 1 1 1 0 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1

[zx8754]

Can you give us sample output?

[kumar.vinod81]

fastPHASE format (2 individuals and 11 SNPs, row wise mutation) ID 5906291001R01C01 1 1 1 0 1 1 1 1 1 0 1 1 1 1 0 1 1 1 1 1 0 1 ID 5906291001R01C02 1 1 1 0 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1

[kumar.vinod81]

It not look like I want to paste here. SNP sites are row wise..

Answer 1

[zx8754]

The output looks like phased data, not imputed data. I suggest to use IMPUTE2

1. Convert plink files to gen format - use GTOOL

2. use IMPUTE2 for imputation

3. then use PLINK to work with dosage output

[kumar.vinod81]

All the site are imputed but I am not able to convert them into PLINK format...I am completely failed.

[kumar.vinod81]

I've not used impute2. Is it easy to use?

Answer 2

[Wen.Huang]

1. convert this format to plink's tped format, which is very similar to fastPHASE outputs, just a few extra columns
2. make a tfam file with the individual ID, etc.
3. with tped and tfam on hands, you can use plink to make various conversion if you need

[kumar.vinod81]

Thanks Huang, but the SNP sites in fastPHASE format are in rows, not in subsequent columns. While in tped format, the SNP sites are in same row and different column. What to do? I am totally puzzled...

[714]

Hi Kumar, did you ever find a solution to this? I am encountering the same problem....

[zx8754]

If you transpose your file, then it is similar to this question: Convert genotype matrix into plink format

[714]

I'm afraid the conversion is not that simple. The genotype for each individual is in rows. To use the example above:

The output of fastPHASE (2 individuals, 11 SNP loci)

ID 5906291001R01C01 1 1 1 0 1 1 1 1 1 0 1 1 1 1 0 1 1 1 1 1 0 1 
ID 5906291001R01C02 1 1 1 0 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1

In short the format goes something like this for 3 SNP loci per individual:

SNP1_Allele_1 SNP2_Allele_1 SNP3_Allele_1 SNP1_Allele_2 SNP2_Allele_2 SNP3_Allele_2

[zx8754]

See gtool -G option for GEN to PED Conversion mode.

[kumar.vinod81]

I didnt get any reply from Biostar but one of my bioinformatician friend written a perl script, and now I am always to send data to him and he send me after its formatting.... But its really not simple with online available tools. Try to contact ur BI team.... Thanks