Consider integrations of NTM reference genomes

[abhi18av]

@Mxrcon here's a dumping ground of ideas regarding this investigation.

[Mxrcon]

CONTEXT:

During our discussions on slack, Hanno was curious to see how mtbseq would interact with the usage of other references. Preferably M. bovis and M. canetii and see if is there any difference from the standard reference: M. tuberculosis h37rv

[Mxrcon]

MTBSEQ manual about using custom refs

Decided to add multiple references to my conda env with mtbseq.

• Standard references:
  • Mycobacterium tuberculosis H37rv (M._tuberculosis_H37Rv_2015-11-13)
  • Mycobacterium abscessus (M._abscessus_CIP-104536T_2014-02-03)
  • Mycobacterium chimaera (M._chimaera_DSM44623_2016-01-28)
  • Mycobacterium fortuitum (M._fortuitum_CT6_2016-01-08)

Every ref has 2 main files: .fasta and genes.txt

• NOTE: M. tuberculosis has additional files:

  • .fasta.bwt
  • .dict
  • .fasta.fai
  • .fasta.pac
  • .fasta.amb
  • .fasta.sa
  • .fasta.ann

• REF seqs used:

  • M. canetii
  • M. bovis

[Mxrcon]

After some hard work, i came up with a python script to generate the genes.txt file, i used the M. Tuberculosis H37rv gbk from ncbi and tried to replicate exactly the output.

script link

I double checked the outputs using the mtbseq source genes.txt and the outputs from my script and running git diff between them, I could replicate exactly the same file so the git diff output was none.

[abhi18av]

CC @jloubser

[jloubser]

@Mxrcon this is fantastic! appreciate the trouble. I've tested the script a couple of times with M.bovis (LT708304), but with no luck. Just for good measure I tried it with H37Rv and worked perfectly. Then I realised I used the genome assembly instead of the complete sequence record. Tried it again, this time with NC_002945. Exactly the same error.

I noticed the Genbank record for H37Rv is linear DNA and for LT708304 is circular DNA and thought it might be the issue. It's not.
I chose this one as this is the reference used by all M.bovis papers (correct name: AF212297).
But I used the BCG reference genome you used and it worked again.
I'll check further what the differences are, but I find this odd.
Any idea why it wouldn't work with these reference sequences?

The error I got with both tries (LT708304 & NC_002945):
<(base) [jloubser@khaos data]$ python3 create_mtbseq_ref.py bovis.gb
Traceback (most recent call last):
File "/home/jloubser/mtbseq-nf/data/create_mtbseq_ref.py", line 40, in
create_fasta(output_name, sample_sequence)
File "/home/jloubser/mtbseq-nf/data/create_mtbseq_ref.py", line 13, in create_fasta
with open(output_name+".fasta", "w") as output_fasta:
FileNotFoundError: [Errno 2] No such file or directory: 'Mycobacterium_tuberculosis_variant_bovis_AF2122/9723-JUN-2022.fasta'>

[Mxrcon]

Hello @jloubser, sorry for my delay. This issue seems to be related to the / character on the file name, would you be able to rename the file name to something like: Mycobacterium_tuberculosis_variant_bovis_AF2122_9723-JUN-2022 and test it again? this should fix the problem

[jloubser]

Hello @jloubser, sorry for my delay. This issue seems to be related to the / character on the file name, would you be able to rename the file name to something like: Mycobacterium_tuberculosis_variant_bovis_AF2122_9723-JUN-2022 and test it again? this should fix the problem

@Mxrcon no problem, you have more pressing things to focus on. I appreciate the time you've put in for this. I'll give this a try.