Taxonomic Classification

Now that we have thoroughly cleaned our data, we are finally ready to assign a taxonomy to our sequences.

We will do this using a Bayesian classifier (via the Classify.seqs tool) and a mothur-formatted training set provided by the Schloss lab based on the RDP (Ribosomal Database Project, Cole et al. 2013) reference taxonomy.

In this tutorial we will use the RDP classifier and reference taxonomy for classification, but there are several different taxonomic assignment algorithms and reference databases available for this purpose.

An overview of different methods is given by Liu et al. 2008 and shown below:

The most commonly used reference taxonomies are:

• SILVA (Quast et al. 2012)

• GreenGenes (DeSantis et al. 2006)

• RDP (Cole et al. 2013)

• NCBI Taxonomy Database (Federhen 2011)

The choice of taxonomic classifier and reference taxonomy can impact downstream results. The figure from Liu et al. 2008 given below shows the taxonomic composition determined when using different classifiers and reference taxonomies, for different primer sets (16S regions).

Figure: Compositions at the phylum level for each of the three datasets: (a) Guerrero Negro mat, (b) Human gut and (c) Mouse gut, using a range of different methods (separate subpanels within each group). The x-axis of each graph shows region sequenced. The y-axis shows abundance as a fraction of the total number of sequences in the community. The legend shows colors for phyla (consistent across graphs).

Which reference taxonomy is best for your experiments depends on a number of factors such as the type of sample and variable region sequenced.

Another discussion about how these different databases compare was described by Balvočiūtė and Huson 2017.

Removal of non-bacterial sequences

Despite all we have done to improve data quality, there may still be more to do: there may be 18S rRNA gene fragments or 16S rRNA from Archaea, chloroplasts, and mitochondria that have survived all the cleaning steps up to this point. We are generally not interested in these sequences and want to remove them from our dataset.

Step 1: Taxonomic Classification and Removal of undesired sequences

This step involves using Classify.seqs with the fasta file and count table generated by Remove.seqs and the reference and taxonomy data that we loaded into this project during the data preparation step. We will then use Remove.lineage we will remove any non-bacterial sequences.