I have received the following files from the 16S rRNA gene sequencing of a set of samples and want to perform taxonomic analysis on them (relative abundance, clearing out unrepresented taxa, removing of taxanomy labels):
- ASVs table (.csv) that contains the ASVs in counts and percentages and a column with taxonomic assignments (shown as taxanomy labels k_, p_, c_ etc)
- Identity table (.csv) that contains the classification of sequences based on a percentage of identity to a reference sequence table. Therefore some counts/percentages are at species level, others at genus or family level.
I'm only new to R and have downloaded the most commonly used packages (phyloseq, microbiome, vegan, microeco). Have you used any simple tutorial on R that covers this type of taxonomic analyses given these input files?
Hello Alessandra,
there are several ways of doing what you want to analyze. I usually work with phyloseq and vegan, although you can try some other packages too.
The first thing to do is import your data into R. As pretty much everything in R, there are many ways of achieving the same ta@sk. I prefer to create a .biom file from the taxonomy table and the count table, that way you can import it into R using phyloseq command "import_biom". Attached you can find how it should look the tab separated file to import it using the terminal in mac or linux environments.
https://biom-format.org/
https://biom-format.org/documentation/biom_conversion.html
Other way of doing it is importing each file separately as a .cvs file. For the taxonomy table you could use the following commands with phyloseq package.
Taxonomy table:
It should contain only the taxonomy and the ASV ID
COG.tax2
Thank you David Rosado-Porto for your answer! Just to be sure before proceeding, if I import those 3 tables you mentioned as comma separated files and then merge them, would I still be able to operate the phyloseq commands you suggested from this page (https://joey711.github.io/phyloseq/import-data.html) or do they need to be in .biom?
The taxonomy column of my ASVs table looks like this:
uncultured prochlorococcus sp. ;k__bacteria;p__cyanobacteria;c__cyanobacteria;o__synechococcales;f__prochloraceae;g__prochlorococcus;s__prochlorococcus spp.; superkingdom__bacteria;
should I keep the taxonomy labels or do I need to convert them into separate columns ("Kingdom", "Phylum", "Order" etc.) when sticking to this same phyloseq tutorial?
Thanx in advance, it was very helpful!
Hello Alessandra Polo
If you are gonna work only with the files in .cvs format you don't need to convert them into .biom. The .biom format is in case you want to have the frequency table and the taxonomy together in the same file. But in your case you can import the three files (Frequency table, Taxonomy table and metadata table) separately and merge them with phyloseq using the commands from my previous comment. One thing to keep in mind, if you're going to use that way, then the Frequency table only should have the ASV ID and their frequency per sample, and the taxonomy table should have only the ASV ID and the taxonomy only (attached examples).
Regarding the taxa labels, you can remove them, as long as each taxa rank is separated from one another phyloseq won't have a problem with it.
Best regards
- Badges
- Science topic
- Similar topics
- Filing