PLS-DA Result Irreproducible

When I change the name of my analyses in the column heading of my CSV file, the PLS-DA result is slightly different compared to when I enter the raw data. What can I do to make it consistent?

The format of the raw data column heading is in the format of “k__Bacteria;p__Gemmatimonadetes;c__Gemmatimonadetes;o__N1423WL;f__;g__;s__” and I would like to change it to A-Z and so on for my 684 different analytes.

It is not very clear to me about your data type - are those features metabolites or microbes?

They are microbes, but I want to statistical analysis should be the same.

There are certain steps unique for microbiome data. Why not use our MicrobiomeAnalyst?

Thank you for letting me know, I had not heard of MicrobiomeAnalyst. I will try it!