search

harry-thorpe / demix_check

GNU General Public License v3.0
2 stars 1 forks source link

selection of reference genomes #2

Open clb21565 opened 1 year ago

clb21565 commented 1 year ago

hi there, a bit confused about the input to this, particularly the --ref setting

are these to be the same genomes used to perform pseudoalignment on, or? are they supposed to be bins?

sorry if i missed it, ty!

clb21565 commented 1 year ago

OK-- in retrospect this might have been a silly question ;-]

a follow up where clarification could be helpful:

before the running of demix_check --mode_setup, should we concatenate the contigs of the individual assemblies as is recommended by mGEMS? or is this built in?

clb21565 commented 1 year ago

OK-- I have it working! yay! putting some details in the hopes that someone will find it helpful:

  1. Need precise versions of mash (I used v2.3 successfully; need mash screen which is not in older versions)
  2. Need specific themisto version (I used v.2.1.0; need the --gzip-output option & the build_index bin)
  3. Need python v >=3.7 (some options used in the subprocess lib used in demix_check were implemented in 3.7)
  4. Do not process the reference genomes prior to running --mode-setup; (I had concatenated contigs + converted to singleline fa – this resulted in an erroneously formatted fasta file/index)
clb21565 commented 1 year ago

another update: i was having issues creating larger sets of reference indices because of memory. adjusted the --mem-megas by editing the reference.py l89; could add the parser argument to demix_check.py instead to make it easier to adjust.

harry-thorpe commented 10 months ago

Hi, I'm really sorry I never replied to these comments - I haven't been actively working on this for ages.

I'm getting back to it now so will take them into account