Annotating genomes with RNA families using Infernal

This tutorial shows how to use the Infernal software to annotate the reference SARS-CoV-2 genome with RNA families from Rfam.

The same approach can be used to find RNA families in any RNA or DNA sequence.

Requirements

• Docker Desktop on your computer in order to access a complete working environment pre-configured using Docker.

💡 Alternatively, try Play with Docker (PWD) in your browser (requires a free Docker account and depends on the resource availability).

Getting started

Download a pre-built Docker image containing all data and software:

docker pull rfam/tutorials

Start an interactive session:

docker run -it rfam/tutorials

You should see a screen similar to the following:

$ docker run -it rfam/tutorials
rfam-user@48a963da2278:~$

You can now type any bash commands and follow the instructions below.

Tutorial

1. Type ls to list files in your folder. You should see:

   • Rfam.cm - Rfam covariance models from release 14.5
   • Rfam.clanin - A list of Rfam clans
   • virus.fasta - SARS-CoV-2 sequence NC_045512.2

2. Run cmpress Rfam.cm to prepare the Rfam covariance models to be used by Infernal (takes ~15 s, you only need to do this once).

3. Run Infernal cmscan to find Rfam families in virus.fasta (the command should take 30-60 seconds):

   cmscan --cut_ga --rfam --nohmmonly --clanin Rfam.clanin --oskip --fmt 2 -o output.txt --tblout table.txt Rfam.cm virus.fasta
   

   Here is a quick explanation of the command line options:

   • --cut_ga - use the thresholds selected by Rfam curators
   • --rfam - run in “fast” mode, the same mode used for Rfam annotation
   • --nohmmonly - run all models in CM mode (not HMM mode). This ensures all GA cutoffs, which were determined in CM mode for each model, are valid
   • --clanin Rfam.clanin --fmt 2 --oskip - remove redundant hits from the same Rfam clan
   • -o output.txt - cmscan output including alignments
   • --tblout table.txt - cmscan output table

   ⚠️ It is recommended to always use the --cut_ga --rfam --nohmmonly options when annotating genomes with Rfam families.

4. Inspect the output files output.txt and table.txt:

   less -S output.txt
   less -S table.txt
   

5. Find the Rfam families from the Infernal output on the figure from Huffsky et al., 2020:

   

6. Bonus points: repeat step 3 without the --oskip option. Notice the additional hits from the bCoV-5UTR and bCoV-3UTR families (see table-no-oskip.txt).

Local development

1. Download or clone this repository and move to the directory:

   git clone https://github.com/Rfam/rfam-tutorials.git
   cd rfam-tutorials
   

2. Build a docker image:

   docker build -t rfam/tutorials .
   

3. Start a docker container and mount the data folder:

   docker run -v `pwd`/data:/home/rfam-user/data -it rfam/tutorials
   

Further reading

• See Alternate Protocol 1 in Kalvari et al., 2018 for more details about annotating a genome with Infernal and Rfam
• Rfam SARS-CoV-2 annotations are described in Huffsky et al., 2020
• Find out about other Infernal commands in the Infernal User Guide

Questions or ideas for improvement?

If you have any feedback, feel free to create an issue or submit a pull request.