This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Input problem with EIGENSTRAT

I'm running EIG 7.2.1 and using convertf to make a geno file with the following parameters:

genotypename:   merged.eigenstrat.34string.length.beded
snpname:    merged.eigenstrat.34string.length.map
indivname:  56001801066929_WGZ.snp.fam
outputformat:   PACKEDANCESTRYMAP
genotypeoutname:    merged.eigenstrat.34string.length.geno
snpoutname: merged.eigenstrat.34string.length.snp
indivoutname:   merged.eigenstrat.34string.length.ind
familynames:    NO

This is only being done for chr22 as a tutorial for myself to generate an easy-to-use wrapper.

I have about 2,504 1000Genomes samples, with one additional unknown sample added in.

The original input is a VCF, which has been modified so that the ID column has only strings < 35 characters long.

the bed file was generated thus:

~/Scripts/plink1.9/plink --vcf merged.eigenstrat.34string.length.vcf --make-bed --out merged.eigenstrat.34string.length

and the map file thus:

~/Scripts/plink1.9/plink --vcf merged.eigenstrat.34string.length.vcf --recode --out merged.eigenstrat.34string.length

The end of the fam file looks like:

Y035    NA19198 0   0   1   -9
Y010    NA18511 0   0   2   -9
Y035    NA19197 0   0   2   -9
Y014    NA18519 0   0   1   -9
56001801066929_WGZ  56001801066929_WGZ  0   0   0   -9

The added sample is the last line. when I run convertf: bin/convertf -p convertf.params.txt

I get a repeat of the input parameters and then the following to STDOUT:

genetic distance set from physical distance
HG01889 ignored
HG01894 ignored
HG01958 ignored
HG02314 ignored
HG01989 ignored
...
NA18511 ignored
NA19197 ignored
NA18519 ignored
56001801066929_WGZ ignored
all individuals set ignore.  Likely input problem (col 6)
resetting all individual...
genotype file processed
numvalidind:   2505  maxmiss: 2505001
packedancestrymap output
##end of convertf run

Why are all individuals being ignored?

How can I modify the input to get rid of the input problem? How can I modify the file so that all individuals are converted? Furthermore, how can I specify which individuals are YRI, CEU, FIN, etc.?

eigenstrat

1 answer

This is an annoying thing about using plink to make eigenstrat input -9 in the last column of the fam file is for missing samples. Change it to 1 for all samples.

Log in to answer this question.