Cerberus

Cerberus is a set of tools designed to characterize and enhance transcriptome annotations. Currently Cerberus can do the following:

• Represent transcript start sites (TSSs) and transcript end sites (TESs) as bed regions rather than single base pair ends
• Integrate intron chains from multiple transcriptome annotations (GTFs) to create a transcriptome of the union of them all
• Integrate TSSs and TESs from multiple GTFs as well as from outside BED sources to create end annotations from the union of them all
• Number intron chains, TSSs, and TESs found by their priority in a reference GTF
• Use the enhanced intron chain and 5'/3' end sets to annotate an existing GTF transcriptome with transcript triplets and to modify the GTF and corresponding abundance matrices to reflect the new naming scheme / identities of the transcripts
• Compute gene triplets for different sets of isoforms for each gene based on the TSSs, ICs, and TESs used among the isoforms of the gene
• Generate plots (see examples below) to visualize gene triplets on the gene structure simplex
• Compute centroids of gene triplet coordinate distributions
• Compute pairwise gene structure simplex distances between pairs of gene triplets

Please visit the Cerberus website for documentation.

Cerberus Docker image

Cerberus' dependencies are sometimes hard to install. For this reason, we provide a Docker image inside which you can run Cerberus. You can also rebuild the image from the Dockerfile.

# pull the image
docker pull ghcr.io/fairliereese/cerberus:latest

# run in interactive mode to have access to Cerberus installation
my_data_dir=<directory with your data to mount>
docker run \
  --rm \
  -it \
  -v ${my_data_dir}/data \
  ghcr.io/fairliereese/cerberus:latest

Once inside the interactive Docker image, activate the environment needed for Cerberus:

conda activate app_env

Following this, you'll directly have access to all CLI Cerberus commands and the Python API.

Cerberus data viewer

We also provide a simplex data viewer which users can launch to explore the TriSET triplets they compute on their data with Cerberus.