search

danielzmbp / lspred

LSpred is a model to predict lifestyle in plant pathogenic organims based on genomic information.
MIT License
0 stars 0 forks source link

Predict lifestyle for a new species #2

Open akiselev-lrsv opened 3 years ago

akiselev-lrsv commented 3 years ago

Dear @danielzmbp , Your work looks amazing, congratulations.

I'm interested in using this tool, could you please briefly explain how I can use to predict a lifestyle for a new species outside your training set? Do I need to assign it to a specific group ad then test the hypothesis, or the software can assign a species to a group automatcally?

Thank you!

danielzmbp commented 3 years ago

@akiselev-lrsv Thanks a lot for your comment! You don't need to assign it to any group, the model will do that for you. You just need to run the model on your annotation using the notebook lspred_testing.ipynb. Note that you need to have the input files in the table format from the genome-properties script. If you have further questions, please don't hesitate to get in touch!

akiselev-lrsv commented 3 years ago

Thank you @danielzmbp , Where can I read about the output scores of the notebook lspred_testing.ipynb? And how to use to interpret the results?

[[0.00992905 0.00214383 0.10016803 0.8877591 ] [0.30474845 0.25972545 0.24817495 0.18735124]] ['saprotroph', 'necrotroph', 'hemibiotroph', 'biotroph']

Best. Andrei

danielzmbp commented 3 years ago

@akiselev-lrsv

Where can I read about the output scores of the notebook lspred_testing.ipynb?

I'm not sure I understand the question. Do you mean what is the source of the examples?

And how to use to interpret the results?

Those are the probabilities for the four output classes. Meaning, the first one would be predicted biotroph, and the second one a saprotroph, but not as clearly as the first one. I will try to make this more explicit in the notebook. Thanks again for your interest, Daniel