I found an error in kneaddata log file. it seems that it wrongly matches the reference database and the decontaminated file. Below is a demo of the issue:
the database is marked as hg37dec_v0.1 but from the file name (SILVA_128fastq), rRNA database is used at this step. This minor error leads to the merged log file wrongly report the number of reads get filtered at each step. It would be very helpful if you could help me to look into this issue.