Potential problem with converting Oxford format to PLINK binary

[Maciej Trzaskowski]

Hi,

I used the new PLINK to convert Oxford format to bed files using:

plink1.9 --gen myGenFile.gen --sample mySampleFile.sample --make-bed --out myNewBedFiles 

It was brought to my attention that the conversion seemed to have shuffled the rs numbers. Whenever we called for some specific SNPs we were getting either the right genotypes but with completely wrong rs numbers attached or the right rs numbers but with the wrong genotypes. We noticed that when converting from Oxford to .ped .map files first and then making binary files resolved that problem. We have run number of tests and it looks like there is a bug somewhere when the direct conversion from Oxford to .bed .bim .fam files happens. 

It would be great to know if that's the case. 

Thanks.

[Christopher Chang]

(i) Does this happen with the 15 Oct stable build?

(ii) If it does, can you provide a small .gen/.sample dataset for me to replicate the bug with so I can try to post a fix later today?  This sounds horrible.

[Maciej Trzaskowski]

Ha, that is a good point. I just checked the version we have on our cluster and it seems to be 1.90a 64-bit (11 Dec 2013)!

I'm afraid I won't be able to instal the updated version until Monday when our cluster administrator is back. I'll ask him to update it and i'll test it again. I will write back with the results. Thank you for getting back to quick Christopher. 

[Maciej Trzaskowski]

Hi Christopher, 

I just tested the latest binaries and they seem to be working fine, so as far as I can see the problem has now been resolved. Sorry for the scare!

Best,

Maciek