Usage of "train.metag" function in melonnpan

Microbial community profiling MelonnPan
1

Hi @himel.mallick - I have trained paired metagenome-metabolome data which resulted into 3 outputs: MelonnPan_Trained_Metabolites.txt
MelonnPan_Training_Summary.txt
MelonnPan_Trained_Weights.txt
For the prediction purpose, the following code was used:

 melonnpan.predict(metag = data_rel_abd, 
                    weight.matrix = "~/Desktop/melonnpan_test/train_output/MelonnPan_Trained_Weights.txt",
                    output = output_directory)

  1. Should I do any change in my code for performing prediction tasks using the model generated from custom (not default) paired metagenome-metabolome data?

  2. I am asking as I was confused about the “train.metag” option of melonn.predict function.

2

Hi @DEEPCHANDA7 - sorry for the late reply. The train.metag option is needed to calculate the RTSI scores which is independent of the prediction task. Your code looks correct from a quick glance.

I hope this helps,
Himel