Open fa8sanger opened 2 months ago
In case this trick works for others: Instead of trying to make uxm work with non-chr hs37d5, I am modifying myself the pat.gz files so they have "chr":
gunzip XXX.pat.gz
perl -pi -e 's/^/chr/' XXX.pat
bgzip XXX.pat
tabix -C -b 2 XXX.pat.gz
./uxm deconv -o kk.csv XXX.pat.gz
This works for me
Dear Netanel and Tommy,
We have some EMseq data mapped to hs37d5, but ours is a version of hs37d5 without "chr" in the chromosome names. This is giving me problems with uxm.
wgbstools bam2pat
works without problems and generates the pat.gz files. However, I haven't managed to makeuxm
work. If I use the downloaded atlases I getWarning: skipping an empty sample
, and if I modify the atlas to remove "chr", then I getInvalid atlas: supplemental/Prueba.tsv. "name" column must all start with "chr"
Is this a known imcompatibility and if so is there any trick? or am I just doing something wrong?
Thank you very much for your work, Fede