About tbox-scan

A lightweight tool for discovering T-box riboswitches in a given FASTA sequence adapted from tbdb.io. Uses INFERNAL for prediction. Employs TBDB feature identification to identify likely specifier sequence. Provides predicted specifier region, most likely specifier, T-box sequence, and makes structure predictions. For transcriptional T-boxes, will also attempt to fold terminator structure. Feature refinement based on gene context and host organism tRNAs is not included in this tool. This tool is part of the T-box Annotation Database (TBDB, https://tbdb.io) collection. Currently only supports input files with a single FASTA header and sequence. Accepts whole genomes as inputs.

Dependencies

This program is written for unix operating systems and requires INFERNAL, python, biopython, conda, and pandas. Installation of dependencies is easiest using conda in a conda environment.

Installation using Conda

You can also easily install tbox-scan as a commnad line tool using conda. Simply use the following conda command.

 conda install -c bioconda tbox-scan

Installation from source (requires pip)

First, download or clone this repo then simply cd into the directory with install.sh. Run the install using sh (e.g. sh install.sh or sudo sh install.sh ) to install tbox-scan. A PATH env will be set to the tbox-scan directory that contains default files for running tbox-scan. The install script will also use pip to install the necessary python scripts to your current python environment. If errors arise, please check for missing python dependencies and correct python environment.

 git clone https://github.com/jamarchand/tbox-scan
 cd tbox-scan
 sh install.sh

Using tbox-scan

Usage: tbox-scan -f <Input FASTA file> [-options]

Scan a fasta sequence file for T-boxes and predict specifier & T-box sequence.
          -- Default: Will use INFERNAL with RFAM00230 covariance model with basic output
          -- Example: tbox-scan  -f input.fa -o output_file.csv -v
Dependencies: INFERNAL, biopython, python3, pandas.


Options
  -f <file>  : input FASTA <file> (required) 
  -o <file>  : save final results in <file> as .csv
                  default: out.csv
  -i <file>  : save INFERNAL output predictions to .txt <file>
                  default: INFERNAL.txt
  -l <file>  : save a .txt log <file> of pipeline output
  -m <model#> : search for t-boxes using specified covariance model
                  1: RFAM model (RF00230.cm), works best on class I t-boxes (default)
                  2: TBDB model (TBDB001.cm), works best on class II t-boxes 
  -c <value> : score cutoff for INFERNAL model predictions (default = 15)
  -v         : save verbose output
  -s         : silence console output
  -h         : print out summary of available options

Examples
    cd examples
    tbox-scan  -f genome_example1.fa -o output_file1.csv -m 1 -v
    tbox-scan  -f genome_example2.fa -o output_file2.csv -m 1 -s -c 100

About this work

Tbox-scan was written as an auxilliary tool for T-box Annotation Database (https://tbdb.io). More information about how the database was built can be found on the BioRxiv page. If you have questions or issues running tbox-scan, post an issue or message me on twitter @syntbio.

• Marchand, J. A., Pierson Smela, M. D., Jordan, T. H. H., Narasimhan, K. & Church, G. M. (2020). TBDB – A database of structurally annotated T-box riboswitch:tRNA pairs. bioRxiv.