search

cdanielmachado / carveme

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

Very different reconstructions with CarveMe 1.5 #126

Closed cdanielmachado closed 3 years ago

cdanielmachado commented 3 years ago

The manual curation applied to the universal model in CarveMe 1.5 seems to result in large changes in reconstructed models compared to previous versions.

This is reported by different users:

https://github.com/cdanielmachado/carveme/issues/125

https://github.com/cdanielmachado/carveme/issues/123

https://github.com/cdanielmachado/carveme/issues/122

I will gather all related issues in this thread while trying to solve this problem.

afvrbanac commented 3 years ago

Just in case this is helpful, I got fairly different models for the same genomes, but they now appear to now work better for anaerobes. I was making models of three anaerobes with previous versions and the models always incorrectly couldn't grow without oxygen. I tried out v 1.5 on the same genomes and the resulting models can all grow anaerobically.

cdanielmachado commented 3 years ago

Thanks for the feedback! It is good to know that at least something is working better. I will try to make sure that improvement doesn't go away when fixing the other problems :)

cdanielmachado commented 3 years ago

Found the problem, it was a bug in the calculation of the score for spontaneous reactions (such as oxygen diffusion). This was excluding oxygen transport from models and forcing the organisms to ferment and reach lower growth rates.

It should be solved now with release 1.5.1.