Analyzing Metagenomic Data with qiime metagenomics

9.3 Pick Operational Taxonomic Units (OTUs)

There are 3 main protocols for OTU picking:

1. 
   pick_de_novo_otus.py: Reads are clustered against one another without any external reference sequence collection. Useful for studying populations where there is poor characterization of existing data. In addition to clustering, the script also performs taxonomy assignment, sequence alignment, and tree-building steps.
   
2. 
   pick_closed_reference_otus.py: Reads are clustered against a reference sequence collection and any reads which do not hit a sequence in the reference sequence collection are excluded. Useful when existing taxa are well characterized and one is not interested in novel ones. Taxonomic assignments in the reference database are assigned to the OTUs.
   
3. 
   pick_open_reference_otus.py: Reads are clustered against a reference sequence collection and any reads which do not hit the reference sequence collection are clustered de novo. Applicable when existing taxa are well characterized and one also wishes to discover novel ones.
   

Example usage (do not enter this command):

pick_open_reference_otus.py –r reference_taxa.fna –i split_library_output/seqs.fna -o otus

-r file path to reference sequences (not required for “pick_denovo” script)

-i file path to input sequences (post-splitting if applicable)

-o name of output directory (this directory will be created when the script is run)

• 
  So let’s go ahead and run the pick_denovo_otus.py script:
  
• 
  Make sure you are still in the qiime directory
  

• 
  pick_de_novo_otus.py -i split_library_output/seqs.fna -o otus
  

1. 
   Pick OTUs: all of the sequences from all of the samples will be clustered into Operational Taxonomic Units (OTUs) based on selected clustering method and user defined sequence similarity threshold. There are many similarity methods already implemented in QIIME, like blast, uclust or usearch.
   
2. 
   Pick a representative sequence set, one sequence, most abundant, per OTU: This representative sequence will be used for taxonomic identification of the OTU and for phylogenetic alignment. The relevant file (seqs_rep_set.fasta) can be found in the otus/rep_set/ directory, along with a log file.
   
3. 
   Align the representative sequence set - necessary if phylogenetic tools such as UniFrac will be used. Alignments can be generated de novo with programs such as MUSCLE, or to an existing reference using tools like PyNAST. The alignment file is created in the directory otus/pynast_aligned_seqs/, along with a log file.
   
4. 
   Assign taxonomies to the representative sequences in the seqs_rep_set.fasta file. The sequences are compared with established databases to define taxonomic identities. QIIME uses the uclust classification system as default. Assigned taxonomies (including confidence values) are written to a text file in the otus/uclust_assigned_taxonomy/ directory.
   
5. 
   Filter the alignment prior to tree building to remove positions that are phylogenetically uninformative. The filtered sequence file is created in the directory otus/pynast_aligned_seqs/.
   
6. 
   Build a phylogenetic tree. Tree building is necessary for UniFrac diversity measurements and other phylogeny based analyses. A Newick format tree is written to rep_set.tre, in the otus/ directory and can be viewed using tree visualization software (for example TreeVector [Pethica, R. and Barker, G. and Kovacs, T. and Gough, J. (2010). "TreeVector: scalable, interactive, phylogenetic trees for the web". PLoS ONE 5: e8934. doi:10.1371/journal.pone.0008934]).
   
7. 
   Build an OTU table: Using the OTU map (step 1) and taxonomic assignments (step 3), QIIME assembles a matrix of OTU abundance in each sample with meaningful taxonomic identifiers for each OTU. The results are written to a file (otu_table.biom) in the otus/ directory.
   

• 
  Let’s examine the results of the pick_denovo_otus.py script. First look at the otu_table.biom file in the otus directory just created by the previous script
  

• 
  We can use another utility, biom summarize-table, to obtain a very brief summary, which lists the number of OTUs identified in each sample (Counts/sample detail):
  

• 
  biom summarize-table –i otus/otu_table.biom –o table_summary.txt
  

• 
  Examine the table_summary.txt output file
  
• 
  Now let’s use summarize_taxa_through_plots.py to group OTUs at different taxonomic levels (phylum, class, order, etc.) and plot the results:
  

• 
  summarize_taxa_through_plots.py -i otus/otu_table.biom \ –o wf_taxa_summary -m Fasting_Map.txt
  

• 
  To view the resulting charts, use SCP/WinSCP to transfer the taxa_summary_plots folder to the desktop of the Mac/PC and open the bar_charts.html file (note: on the Mac, you will need to use the –r flag to transfer the whole folder):
  

• 
  scp –r taxa_summary_plots user.name@csurs11.csr.uky.edu: