Downloading ribosomal rna fasta files

In order to understand the taxonomic composition of metagenomic samples it is need to classify sequences against a reference database. However, obtaining and using a database for taxonomic classification isn't trivial and the difficulties can be due to a range of issues you could encounter. These issues could be due to the sequences you want to classify (is it shotgun data, or is it a marker gene like 16S rRNA, rpoB or another gene), the sequence repository you are using, the taxonomic content of the database, or your classification tool wants to have sequence data in a special format. That means that it is important to obtain a database that is suitable for your data, that has good taxonomic coverage of the taxa you are interested in, and that it is well curated (e.g. contains few misclassifications).

For instance the large NCBI NT or NR database contain sequences from most studied taxa, and when used for shotgun sequence it is able to classify the largest amount of sequences compared to most other databases. However, the taxonomic classification of many sequences using these two database can be problematic due to contamination, misclassifications or simply because people have been sloppy when adding their data to this repositories. Because of that, many initiatives were started to generate curated databases that had a specific audience in mind.

For instance ribosomal RNA sequences are highly abundant in the NCBI NT database, but their classification in that database is problematic and the quality of many sequences is poor. Several initiatives exist that have dealt with these issues and that resulted in different rRNA databases such as Greengenes, Ribosomal database Project, and the SILVA database. Each of these databases come with their own curation method applied to the rRNA data (imported from NCBI NT database), which results in slightly different databases.

My own preference for 16S rRNA database is to use the SILVA database (lucky for me it is considered to be the better database by several people in the field (what ever that means)).

Acknowledgements

In this tutorial we will follow the tutorial created by on how to obtain the latest complete SILVA database (version ) and make it suitable for use with the mothur pipeline. His tutorial is found here: Building SILVA reference files.

The SILVA database

Before we do anything, take your webbrowser and surf to the SILVA database homepage. We first will explore this resources before starting with the good stuff. You will see something like the image below:

If you are interested in obtaining sequences you can use several options on the SILVA portal. Those are: , , and . Note that the SILVA database contains sequence data for the small sub unit (SSU) and the large sub unit (LSU) of the ribosome found in prokaryotes and eukaryotes. Those are the 16S / 18S rRNA (SSU) or the 23S / 25S (LSU) rRNA molecules. The SSU database is the larger of the two and contains most sequences, but in principle both can be used. Remember that it is only due to history that most microbiome studies use the 16S rRNA, but that the marker is not the most optimal for classification.

Okay let's check the menu options at the SILVA homepage. The and options allow you to collect sequences belonging to taxa of interested. For instance you can collect all sequences belonging to the or only those from the . When you use browse you can collect all those sequences.

In the browser window:

When you use the search option you can filter based on sequence quality. Search sequences with the following options:

• Taxonomy = FirmicutesSee
• Sequence length >
• Sequence quality > 95
• pintail quality > 90

The aligner is good when you have an rRNA sequence and want to know what is actually is, or if you want to obtain reference sequences to build a taxonomy.

Here is a set of Thermosipho sequences that you can use to classify and obtain reference sequences for further use. Download this file to your computer and then upload it to the SINA aligment page. After selecting the file, click the box . This will allow you to select the number of neighbors per query sequences (max = ). You can also change the minimum similarity of those hits with your query sequences. Then run the aligner tool. After it has finished you can see the classifications and you can add neighbor sequences to your shopping cart.

This was a short introduction to the SILVA database homepage. You do not often need it, but for some questions, it is good to know that it exists and that it can be a good resource in case you need to build phylogenies based on an rRNA molecule.

Now we will start with building the database locally so we can use it for our own amplicon analysis. The first thing we need to do is login into our biolinux machine as the . The reason for that is that we need to create a file which will export the silva_db from the arb format into a mothur compatible fasta file.

Adding the mothur export file to the biolinux arb installation.

In order to export the arb database sequence we want to export them with the mothur formatting. Open a new terminal and type:

In this directory we find the files with the export formats for arb. We will create the file: .

Inside nano type:

Downloading the SILVA database

Now that we have created this file, we will change to our normal user account on the biolinx machine.

• Open a terminal .
• go to your Desktop directory and create a folder called: .
• change into the directory .

Now it is time to download the SILVA database. When you take a look at the menu of the SILVA homepage, you will find three options: , and .

We when we just wanted to download a fasta file than we could be interested in the , which contains many folders.

For instance, the latest version of the Silva database, which is release , can be found under the current link folder. There you find a large number of files. The file explains the difference between the files.

However, we want to make the sequences from the ARB database compatible with our own sequences. So we need to ARB files. The are needed for the ARB software, which is the software in which the SILVA database is maintained. We will be using that software on our biolinux since it is installed there.. For more details on ARB see here: www.cronistalascolonias.com.ar

We are going to download an ARB file with the SILVA database. It will not be the complete database, but a non-redundant database, where 16s rRNA sequences are clustered with 99% sequence similarity cut-off. That compressed file is still ≈ Mb big. The commands we need to download it are:

Now we can start up ARB with this ARB database with the command:

Exporting the database to a fasta file

This database contains almost sequences and can not be used yet for mothur. We need to filter out low quality sequences and chimeras and then we need to export the data to a fasta file. When ARB is running do the following:

1. Click the search button

2. Set the first search field to ‘ARB_color’ and set it to www.cronistalascolonias.com.ar on the equal sign until it indicates not equal (this removes low quality reads and chimeras)

3. Click ‘Search’. This yielded , hits

4. Click the “Mark Listed Unmark Rest” button

5. Close the “Search and Query” box

6. Now click on File->export->export to external format

7. In this box the Export option should be set to marked, Filter to none, and Compression should be set to no.

8. In the field for Choose an output file name make sure the path has you in the correct working directory and enter: .

9. Select a format: . This is a custom formatting file that Pat Schloss has created that includes the sequences accession number and it’s taxonomy across the top line. We create this file above, remember?

10. Save this as www.cronistalascolonias.com.ar_vfasta (This will take about minutes and creates a huge (≈ 30 Gb) fastafile). This file should contain .

11. You can now quit ARB. (When your biolinux machine is reacting slowly, restart the virtual machine down after closing ARB). That will clear the memory.

Screening the sequences

Now we need to screen the sequences for those that span the 27f and r primer region, have 5 or fewer ambiguous base calls, and that are unique. We’ll also extract the taxonomic information from the header line. Run the following commands from a bash terminal:

The mothur commands above do several things. First the command removes sequences that are not full length and have more than 5 ambiguous base calls. Note: this will remove a number of Archaea since the ARB RN reference database lets in shorter (> bp) archaeal 16S rRNA gene sequences. Second, www.cronistalascolonias.com.ar convert any base calls that occur before position and after to to make them only span the region between the 27f and r priming sites. Finally, it is possible that weird things happen in the alignments and so we unalign the sequences (www.cronistalascolonias.com.ar) and identify the unique sequences (www.cronistalascolonias.com.ar). How many sequences do we end up with? We can check that:

The last command gives us the following table:

So the filtering reduced the number of sequences to Let's continue with the commands on the bash commandline. Next we convert the resulting fasta file into an accnos file and then we use mothur to pull out the unique sequences from the aligned file (). The following commands are all run from the normal commandline.

At this point we have a full database file called: , and we have a file with the taxonomy extracted from the fasta headers of the align file: . This still contain sequences. But we are not yet ready to use this in mothur.

Formatting the taxonomy files

Now we want to make sure the taxonomy file is properly formatted for use with mothur. First we want to grab the SILVA taxa mapping file by running the following in bash:

Thanks to Eric Collins at the University of Alaska Fairbanks, we have some nice R code to map all of the taxa names to the six Linnean levels (kingdom, phylum, class, order, family, and genus). We’ll run the following code from within R:

Building the SEED references

The first thing to note is that SILVA does not release their SEED; it is private. By screening through the ARB databases we can attempt to recreate it. Our previous publications show that www.cronistalascolonias.com.ar with the recreated SEED does an excellent job of realigning sequences to look like they would if you used SINA and the true SEED. Now we want to try to figure out which sequences are part of the seed. Earlier, when we exported the sequences from ARB, we included the align_ident_slv field from the database in our output. Let’s generate an accnos file that contains the names of the sequences with % to the SEED database and then use mothur to generate SEED fasta and taxonomy files. While we’re at it we’ll also generate the nr_ taxonomy file as well. The following code will be run from within a bash terminal:

The accnos file should now contain a list of sequence IDs. Do you remember how you could count the items in such a list on the commandline?

So now we are done with building the SILVA database

Taxonomic representation

When you are interested in knowing the difference between the full SILVA database and the seed SILVA database than I advice you to run the code that you can find on the tutorial page from Pat Schloss: Building SILVA reference files

archiving the silva_databases

The final step in this tutorial is to archive it so you can store it somewhere safely

the commands:

store it safely, for example.

The database file created for this tutorial can also be found in the nnK directory. Specifically in the directory:

When you want to use those, copy them to your own directory and uncompress the files with:

Now you should be set to classify your own SSU rRNA sequences.