ersilia-os / eos24jm

GNU General Public License v3.0
0 stars 0 forks source link

BioModel Discussion Issue for eos24jm #6

Open Zainab-ik opened 2 months ago

Zainab-ik commented 2 months ago

Summary !!!

GemmaTuron commented 2 months ago

Is QcrB a protein complex? Yes, it is a cytochrome complex, the target of the model

For this model, there'd be 3 outputs; actively inhibits QcrB, moderately inhibits and doesn't inhibit at all. What should best describe the metadata output; growth inhibition or the activity classes? That is a good question, I don not have a clear answer, I guess Active Inactive would suffice?

8 ML algorithm was built for this prediction, the ecfp6-SVM is the best model, and from the source repository, that's the model deployed. I'd be adding only SVM as the metadata, right? Yes, exactly

GemmaTuron commented 2 months ago

The metadata looks good! Is there any validation in the lab for the model?

Zainab-ik commented 2 months ago

The metadata looks good! Is there any validation in the lab for the model?

There's no experimental validation.

GemmaTuron commented 2 months ago

Please @Zainab-ik double check that all the metrics listed are used

Zainab-ik commented 2 months ago

Please @Zainab-ik double check that all the metrics listed are used

All metrics are used.

GemmaTuron commented 2 months ago

then let's add it to biomodels

Zainab-ik commented 2 months ago

then let's add it to biomodels

On it.