plink 1.9 --homozyg-group

[ber...@sas.upenn.edu]

Hi plink users and devs,

I think I encountered a bug in 1.9 when identifying runs of homozygosity using the --homozyg --homozyg-group flags. At least one run of homozygosity is recorded in the .hom file, but not in the .hom.overlap file. It very much passes the criteria for being an RoH, and it does appear in the .hom file. I was wondering if there is an explanation for this (it is the only sample in the set that has an RoH in that region, so maybe the .overlap file doesn't include it if it doesn't have an overlap?) or if it is indeed a bug. The command seems to have worked really well other than a missing just a few runs from the .overlap file.

Here is the exact command I ran:

./plink --file test1 --noweb --map3 --autosome-num 64 --homozyg --homozyg-group --allow-no-sex --homozyg-snp 20 --homozyg-kb 10 --homozyg-window-snp 25 --homozyg-window-missing 1 --homozyg-window-threshold 0.04 --homozyg-density 50 --homozyg-window-het 0 --out test_out

Thanks!

Alex

[ber...@sas.upenn.edu]

Here is the log:

PLINK v1.90b4.6 64-bit (15 Aug 2017)

Options in effect:

  --allow-no-sex

  --autosome-num 1

  --file 142_samples_filter04_09.12

  --homozyg

  --homozyg-density 50

  --homozyg-group

  --homozyg-kb 10

  --homozyg-snp 20

  --homozyg-window-het 0

  --homozyg-window-missing 1

  --homozyg-window-snp 25

  --homozyg-window-threshold 0.04

  --noweb

  --out 142_samples_filter04_09.12_out

Hostname: node107.hpc.local

Working directory: /gpfs/fs02/home/berrya/assemblies/plink/plink_1.9

Start time: Tue Sep 12 13:19:00 2017

Note: --homozyg-group flag deprecated.  Use '--homozyg group'.

Note: --noweb has no effect since no web check is implemented yet.

Random number seed: 1505236740

193668 MB RAM detected; reserving 96834 MB for main workspace.

Scanning .ped file... done.

Performing single-pass .bed write (28733 variants, 123 samples).

--file: 142_samples_filter04_09.12_out-temporary.bed +

142_samples_filter04_09.12_out-temporary.bim +

142_samples_filter04_09.12_out-temporary.fam written.

28733 variants loaded from .bim file.

123 samples (0 males, 0 females, 123 ambiguous) loaded from .fam.

Ambiguous sex IDs written to 142_samples_filter04_09.12_out.nosex .

Using 1 thread (no multithreaded calculations invoked).

Before main variant filters, 123 founders and 0 nonfounders present.

Calculating allele frequencies... done.

28733 variants and 123 samples pass filters and QC.

Note: No phenotypes present.

--homozyg: Scan complete, found 39 ROH.

Results saved to 142_samples_filter04_09.12_out.hom +

142_samples_filter04_09.12_out.hom.indiv +

142_samples_filter04_09.12_out.hom.summary .

2 pools of overlapping ROH present.

ROH pool report written to 142_samples_filter04_09.12_out.hom.overlap .

[Christopher Chang]

As the name suggests, the .hom.overlap file only contains overlapping RoHs; any run which does not overlap with a run from another sample is intentionally excluded.