Thank You Mike.
I followed the same workflow for RESCRIPt as is mentioned here and just changed 138.1 to 132 (I used complete set of reads and did not extract them out based on the primers) but as I explained, it did not work for the 18S data.
Further, my biggest suspicion arises from the size of the trained classifier which is way smaller than the size of SILVA 132 pre-trained. Is it possible that I could share my workflow and if possible some artifact or visualization files, if that becomes of any help?
Thank You