Fewer tokens than expected error in Plink when using --filter

Question

[tz.rzezniczak]

Hello,

I am working with .bed .bim and .fam files from the Drosophila Genetic Reference Panel. I only used a subset of the 205 individuals in the dataset and am trying to filter these out in PLINK using the --filter comand, however I keep getting an error, indicating that my .txt file which specifies which individuals to filter has fewer tokens then expected. Can anyone help in identifying what is causing this error?

plink --bfile dgrp2 --filter incl_dgrplines.txt 1 --no-pheno --make-bed --out dgrp3

PLINK v1.90b3.43 64-bit (13 Nov 2016)      https://www.cog-genomics.org/plink2
(C) 2005-2016 Shaun Purcell, Christopher Chang   GNU General Public License v3
Logging to dgrp3.log.
Options in effect:
  --bfile dgrp2
  --filter incl_dgrplines.txt 1
  --make-bed
  --no-pheno
  --out dgrp3

3882 MB RAM detected; reserving 1941 MB for main workspace.
4438427 variants loaded from .bim file.
205 people (205 males, 0 females) loaded from .fam.
Error: Line 1 of --filter file has fewer tokens than expected.

Answer 1

[chrchang523]

For the record, the problem here was that the --filter file used UTF-16 encoding, which plink doesn't understand. Conversion to plain ASCII/UTF-8 without a byte-order mark solves the problem.