fgvieira/ngsLD

ngsld pruning running forever

HomereAMK opened this issue · 1 comments

Hello Filipe,

I got this script that has been running forever now (142hours now).
the input is a beagle file coming from low-coverage whole-genome sequencing

#PBS -l nodes=2:ppn=40
#PBS -l walltime=100:00:00
#PBS -l mem=220gb
#Run ngsLD

perl /services/tools/ngstools/20190624/ngsLD/scripts/prune_graph.pl
--in_file $LDFILES
--max_kb_dist 2000
--min_weight 0.5
--out $OUTPUTFOLDER/xxx_unlinked2.id

I get an output with only the SNPs positions but not the LD estimation...
head ../$OUTPUTFOLDER/xxx_unlinked2.id
scaffold10:10032428
scaffold10:10041873
scaffold10:10042064
scaffold10:10042614
scaffold10:10059026
scaffold10:1006231
scaffold10:10080234
scaffold10:10085198
scaffold10:1009028
scaffold10:10090757

Is it because it's low cov?
All the best,
Homère

Hi Homere,

from what I understand, you are running the prune_graph.pl script with a beagle file?

If so, you need to run ngsLD with the beagle file to calculate LD between pairs of SNPs, and only then use the script prune_graph.pl to prune linked sites.