Unassigner

Evaluate consistency with named bacterial species for short 16S rRNA marker gene sequences.

Summary

The 16S rRNA gene is found in all bacteria, and its gene sequence is highly conserved. Amplification and sequencing of bacterial 16S rRNA genes is a common method used to survey bacterial communities in microbiome research. However, high throughput instruments are unable to sequence the entire gene. Therefore, a short region of the gene is selected for amplification and sequencing.

The resultant sequences, spanning part of the 16S gene, can be used to identify the types of bacteria present in a specimen. For example, one sequence might be assigned to the Streptococcus genus based on sequence similarity. Many programs are available to carry out such taxonomic assignment.

It is generally thought that the 16S rRNA gene is not suitable for assignment of bacterial species. We agree, but with a catch: the gene sequence is suitable for ruling out assignment to many bacterial species. This software is designed to rule out all the species designations that are inconsistent with a partial 16S rRNA gene sequence. For those species that are not definitively ruled out, we assign a probability that the sequence is inconsistent with the species.

Because the software is geared towards ruling out species rather than deciding on the best assignment, we call it the unassigner. It's a cheesy joke, but we've decided to roll with it.

The unassigner library provides a command-line program, unassign, that takes a FASTA file of DNA sequences in a 16S gene region, and gives the probability that the sequence is inconsistent with nearby bacterial species.

Installation

Install with conda using:

conda create --name unassigner -c conda-forge -c bioconda unassigner

Or run with Docker using:

docker run --rm -it ctbushman/unassigner:latest unassign --help

Alternative Installation

Unassigner can be installed using pip:

pip install unassigner

But will require vsearch to be installed separately. It can also be installed from GitHub:

git clone https://github.com/PennChopMicrobiomeProgram/unassigner.git
pip install unassigner/

Usage

The unassign program requires one argument, a FASTA-formatted file of short 16S sequences:

unassign my_sequences.fasta

If the program has not been run before, it will automatically download the bacterial species data it needs, format its reference files, create an output directory named my_sequences_unassigned, and write a table of results there, along with some auxiliary output files. Note that the output directory will be in the same directory as my_sequences.fasta.

Please see the output of unassign --help for a list of the available options.

Trim ragged

The trimragged program takes in a query sequence to search and trim and an input fasta file (or it can read from stdin):

trimragged AGAGTTTGATCCTGGCTCAG --input_file my_sequences.fasta

Trimragged is included to extract different regions from the full length 16S rRNA gene. The purpose of this auxiliary software is to account for the full length 16S rRNA sequences where only a part of the primer is present in the sequence. This can be due to low quality at the beginning or at the end of a sequence due to limitations of sequencing platforms.

The software operates in three steps: 1) Matching the full length of the primer, 2) Matching the partial primer, 3) Aligning reads to other sequences with a known primer location. The sequence of the primer to search and trim is required for the software. Only one primer is accepted at a time, so the user needs to run the software twice with each primer sequence.

Step 1: The software first searches for the full length of the primer sequence. If mismatches are allowed, then the software expands all possibilities of the primer sequence mutations in a list and searches for each. Once a hit is found, the start and end index is stored as a PrimerMatch object.

Step 2: If the min_partial argument is greater than 0, the software then searches for partial matches of the primer in the remaining sequences. The software makes a list of all the possibilities of primers, removing nucleotides from the beginning of the sequence till the minimum length specified by min_partial is reached. Then the software searches for each of the possible primer sequences. Once a hit is found, the start and end index is stored as a Primer Match object.

Step 3: The last part of the software relies on building a database of the sequences with already identified primer sequences from the previous two steps. Then the rest of the reads are aligned against the database of sequences with known primer locations using vsearch. Once a hit is found, and the positions of the primers are estimated by extending the aligned region.

Please see the output of trimragged --help for a list of the available options.

Contributing

We welcome ideas from our users about how to improve this software. Please open an issue if you have a question or would like to suggest a feature.