m2m recon not working with gbff file inputs

AuReMe / metage2metabo

From annotated genomes to metabolic screening in large scale microbiotas

https://metage2metabo.readthedocs.io

GNU Lesser General Public License v3.0

53 stars 7 forks source link

m2m recon not working with gbff file inputs #43

Closed alsmadin01 closed 1 year ago

alsmadin01 commented 2 years ago

Hi @ArnaudBelcour,

I am trying to use .gbff files downloaded from NCBI for metabolic reconstruction using Pathway Tools (m2m recon). However, I am obtaining the error attached. Please note that m2m recon works for me when using .gbk files as input.

Thank you.

Best regards,

Noor Alsmadi

ArnaudBelcour commented 2 years ago

Hi @alsmadin01,

The error associated with your log is due to mpwt not showing correctly that the run of PathoLogic on DEUA01000019 has failed. I have fixed this in mpwt version 0.8.1. By updating to this version, mpwt should show that there is an error with this genome.

And indeed, the gbff associated with this genome and the corresponding genome on the NCBI (DEUA01000019) do not contain any genes or proteins. So Pathway Tools can not predict metabolic networks from these data as it does not perform annotation of the genomes. So you have to predict the genes and the functional annotation of these genes for this genome with annotation tools (such as prokka , bakta or eggnog-mapper).

Best regards, Arnaud Belcour.

alsmadin01 commented 2 years ago

Hi @ArnaudBelcour,

Thank you for your response. Yes – I did my annotations with Prokka and M2M now works. I also have a question regarding the seeds file for M2M: I have a list of micronutrients, minerals, and amino acids – along with their concentrations – that I want to use to create a seeds sbml file (https://insights.envigo.com/hubfs/resources/data-sheets/2018sx-datasheet-0915.pdf). Any recommendations on how to do that?

Best regards,

Noor Alsmadi

cfrioux commented 2 years ago

You can do it with m2m seeds (doc) that will create a seed SBML file from a simple file with a list of nutrients identifiers (the way they look in the SBML file of your reconstructed metabolic networks, i.e with their compartment). A few remarks on that:

You still need to turn your medium content into a list of IDs matching the database you use (e.g. metacyc). This is unfortunately manual
I advise that you put the cytosol compartment for the seeds so that the absence of transport reactions due to the automatic reconstruction does not impede the simulations.
You might also test adding a few cofactors (M_ADP_c, M_NAD_c, M_NADP_c in Metacyc) to prevent any missing activated functions due to the incomplete activation of cycles (once again, because of the automatic reconstruction).

cfrioux commented 1 year ago

I'll close this issue now. Reopen it if needed.