Otu_table(ps) normalization for running melonnpan.predict() function

Hi, I am new in programming. As I see “metag” argument of the melonnpan.predict() only accepts proportional data ranging from 0 to 1.0 , can you please let me know how can I transform my otu_table into an acceptable format for this function? My samples IDs are in rows and otu sequences are columns and otu_table is a matrix. Thank you

Hi @saif - unfortunately, the melonnpan.predict() function can only take UniRef90 as input. We currently don’t support non-UniRef90 input for this function. So in your case, an OTU table as input is not going to work.

Hi @himel.mallick, thank you for your response. For another thing - I was trying to run melonnpan.predict() and melonnpan.train() using the data attached with the R package. However, it was not saving any output in my working directory. For melonnpan.predict(), should not it save a txt/ csv file in the path given to the “output” argument? And for melonnpan.train() function, I was expecting to see the default files mentioned for the outputString argument: ‘MelonnPan_Training_Summary’, ‘MelonnPan_Trained_Weights’, and ‘MelonnPan_Trained_Metabolites’. Could you please explain why I do not see any output files in my working directory and suggest me a way how to get the output files?

I ran the following codes:

melonnpan.predict(metag, output = “D:/BaseSpace”)
melonnpan.train(metab, metag)

Thank you

Hi @saif - I can’t think of a reason not to see the expected output files if the code has been successfully executed. Can you share the log messages?

Hi @himel.mallick, Thank you. I have solved the problem.