EIGENSOFT: CONVERTF ignoring samples

Question

[aritra90]

Hi, the CONVERTF program in EIGENSOFT package is ignoring few samples and not calculating their principal components.

I am seeing this in my terminal when executing CONVERTF.

genetic distance set from physical distance

Sample1 ignored
Sample2 ignored
Sample3 ignored

.....................

Samplen ignored

genotype file processed
numvalidind:     21  maxmiss: 21001
packedancestrymap output
##end of convertf run

Usually, it says ignoring all samples and then outputs the principal components for all the samples. I am just inputting the .PED and .MAP files in par.PED.EIGENSTRAT. It'd be great if anybody can give me some insight on this.

Thanks,
Aritra

Answer 1

[Philipp Bayer]

I've ran into this one before - the EIGENSOFT FAQ has it, too:

Question: convertf decides to “ignore” all my samples. Why?
Answer: A likely reason is that you are using a “fam” or “ped” file with a funny value (0, 9 or -9) in column 6. Try setting column 6 to 1.

(Column 6 is the phenotype as far as I remember, and it has no effect on the PCA results)

[aritra90]

Yes, I saw the same just right after posting :P

Thanks, Philipp, it works now!

[Phoenix Mu]

Although the answer says that "Try setting column 6 to 1." I assume that it does not influence PCA results at all?

[Philipp Bayer]

If I remember correctly, that column is reserved for phenotypes, which are not used in the PCA