Metaphlan3 classification - comparison of samples and RiboDetector

I have a set of meta-RNA-Seq fastq data from bacterial communities. They should be depleted for rRNA, but I want to test the performance of the depletion step.

I have analyzed two paired-end data (#1769, #1770) from the meta-RNA-Seq dataset and compared it to RNA-Seq data (from poly-A sequencing, #516) and to 16S Amplicon-Seq data (#324).
The amount of rRNA reads in the meta-RNA-seq data are ca. 20-30% (analyzed by RiboDetector or SortMeRNA)

However, I was not able to assign rRNA reads from one of the two meta-RNA-Seq data (#1769) to any bacterial taxa.
However, this is also true for the RNA-Seq data (ok, only 3% is rRNA), but alos for the 16S amplicon-seq data (almost exclusively rRNA reads).

Is there an explanation for this. I was searching for optional parameters related to taxonomic classification, but could not find anlything.

The command was: metaphlan (FILENAME).fastq.gz --input_type fastq -o profiled_(FILENAME).txt --nproc 24 --bowtie2db PATH_TO_BT2-INDEX/metaphlan_bt2/ --unknown_estimation

Dear @ARW-UBT , MetaPhlAn doesn’t use rRNA data for taxonomic classification, instead it uses species-specific marker genes (genes exclusive of each species). If you want to check the performance of your depletion step, you should look at tools that classify based on rRNA sequences (e.g. QIIME).

Thanks for you comment. Seems that I was misled by the qiime2 metaphlan plugin, thus assuming that qiime2 classification tools would be used by metaphan, too.

Best regards,

1 Like