MILP solver failed: integer infeasible

bcoltman commented 4 years ago

Hello Daniel, I know Carve was not first developed with eukaryotic model generation but I have been trying to use the iFungi yeast model from https://github.com/LMSE/fyrment/tree/master/genre/xml/01_kingdom as a universal model to generate a yeast strain model from a DNA fasta. The model is aligning and scoring reactions correctly, identifying approximately ~730 reactions. Unfortunately, the generation failed due to an MILP solver failure. I have attached the output as well as the carve call and have also got the debug files identifying reaction match. I have attempted to look at the lp output but I am not too sure where to go with it/bring it back into the carve pipeline. I would appreciate your help/highlighting of any errors I am making!

$ ./20200317_0001_carveme
Running diamond...
diamond blastx -d /home/bcoltman/anaconda3/envs/metabolic_modelling/lib/python3.7/site-packages/carveme/data/input/bigg_proteins.dmnd -q 0001.fa -o 0001.tsv --more-sensitive --top 10
diamond v0.9.30.131 (C) Max Planck Society for the Advancement of Science
Documentation, support and updates available at http://www.diamondsearch.org

#CPU threads: 12
Scoring parameters: (Matrix=BLOSUM62 Lambda=0.267 K=0.041 Penalties=11/1)
Temporary directory: 
Opening the database...  [0s]
Percentage range of top alignment score to report hits: 10
Reference = /home/bcoltman/anaconda3/envs/metabolic_modelling/lib/python3.7/site-packages/carveme/data/input/bigg_proteins.dmnd
Sequences = 84674
Letters = 32226457
Block size = 2000000000
Opening the input file...  [0s]
Opening the output file...  [0s]
....
Closing the input file...  [0s]
Closing the output file...  [0s]
Closing the database file...  [0s]
Deallocating taxonomy...  [0s]
Total time = 11.548s
Reported 7226 pairwise alignments, 7229 HSPs.
1255 queries aligned.
Loading universe model...
Loading reference model...
Loading media library...
Scoring reactions...
Reconstructing a single model
MILP solver failed: integer infeasible
Done.

20200317_0001_carveme
#!/bin/bash
carve --dna 0001_CDS.fa  --gapfill MeOH,Gluc,Glyc --mediadb 0001.tsv \
 --universe-file iFungi_redefined.xml  --fbc2 -o 0001.xml\
 --soft constraints.tsv \
 -d -v

cdanielmachado commented 4 years ago

Maybe there are some blocked reactions in your yeast model that make the problem infeasible.

Can you try running your model through memote? https://memote.io/

bcoltman commented 4 years ago

Apologies, I really should have ran it through MEMOTE before submitting the issue. Around 50% of all reactions were identified as being blocked; thus, it is very likely that is the problem. Thanks for your help, I will reopen the issue if I fix the model and get any further problems.

Rachita-Kumar commented 1 month ago

Hello, I am currently trying to reconstruct GEMs from eukaryotic MAGs using the same universal template, and I encountered similar issues when I ran memote. I was wondering if you have found a solution or strategy to overcome the blocked reactions, or if you used a different template for your analyses? I would really appreciate any help and insights you may have, thanks in advance!

cdanielmachado / carveme

MILP solver failed: integer infeasible #70