search

cdanielmachado / carveme

CarveMe: genome-scale metabolic model reconstruction
Other
145 stars 49 forks source link

inconsistencies and unconserved metabolites in MEMOTE results for carveme models #152

Closed shumantov closed 2 years ago

shumantov commented 2 years ago

Hi Daniel,

I have used carveme to built multiple GEMs, and ran them through MEMOTE to asses their quality. For the GEMs I have tested, stiochiometric consistency is 0 and 6 unconserved metabolites keep show up:

["h2_e","h2_p","h2_c","h_p","h_c","h_e"]

I know this might be a MEMOTE issue, but after opening it over there they suggested to raise it here, since there might a problem with carveme. Not completely sure I deeply understand these results, other than that the models might be inconsistent and not feasible, and I would highly appreciate any insight regarding them.

here are 2 MEMOTE reports:

https://memote.io/report/328e027c-4e0d-4138-a8e7-1cf05f812429 https://memote.io/report/28ec9ebe-a305-4eb5-ac7b-bdbd8cf18e5c

The general script I'm using for creating the models is such:

carve path/bin.219.faa --fbc2 -o path/bin_219_model.xml

Thank you in advance!

Best regards, Alon

cdanielmachado commented 2 years ago

Hi Alon,

That's correct. Unfortunately the BiGG database, where we get the reactions and metabolites from, contain reactions with unbalanced metabolites. We fix or remove most of them, with the exception of proton balance, which is just too widespread to fix consistently.

We haven't observed any situations where this leads to incorrect simulations though.

shumantov commented 2 years ago

I understand. Thank you!