Happie is a wrapper for several independent programs used to identify different mobile elements. It is designed to thrive on HPC environments taking advantage of either Docker or Singularity to containerize the different programs.
Happie allows you to identify and extract all the "mobile" regions of a genome assembly -- those identified as being either prophage, plasmid, or genomic island sequence. This can be used to assess the charactaristics of the mobilome as opposed to the whole genome.
In the next few months, we will be uploaded a few case studies showing happie's utility.
click here for a quickstart guide!
Happie can manually be installed from github as follows
git clone https://github.com/nickp60/happie
cd happie
conda create -n happie biopython pyyaml
source activate happie
python setup.py install
happie_install_or_update
That is going to take care of downloading Docker images for ProphET, prokka, mlplasmids, dimob, mobsuite, PlasFlow, and other goodies. If you are running singularity, a folder called .happie will be created in your home directory.
(Optional) get some test data:
bash ./get_test_data.sh
And try running it!
happie --contigs ./test_data/ecoli104.fna --output tmpresults
Or, if you are using singularity:
happie --contigs ./test_data/ecoli104.fna --output tmpresults --virtualization singularity
You should get several output files:
tmp__small/
├── contig_names_key
├── happie_args.yaml
├── intermediate_files
│ ├── ProphET
│ ├── QC
│ ├── cgview.tab
│ ├── dimob
│ ├── mlplasmids
│ ├── mobile_abricate
│ ├── mobile_annofilt
│ ├── mobile_prokka
│ ├── wgs_abricate
│ ├── wgs_annofilt
│ └── wgs_prokka
├── results
│ ├── mobile_abricate.tab
│ ├── mobile_genome_coords
│ ├── mobile_test.fasta
│ ├── mobile_test.gbk
│ ├── mobile_test.gff
│ ├── wgs_abricate.tab
│ ├── wgs_test.gbk
│ └── wgs_test.gff
└── sublogs
├── QC.log
├── mobile_PropheET.log
├── mobile_abricate_resfinder_log.txt
├── mobile_abricate_vfdb_log.txt
├── mobile_annofilt.log
├── mobile_mlplasmids.log
├── mobile_prokka.log
├── wgs_abricate_resfinder_log.txt
├── wgs_abricate_vfdb_log.txt
├── wgs_annofilt.log
└── wgs_prokka.log
Happie does the following things:
Happie removes short contigs and renames the contig names (to avoid issues with too-long names, non-standatd characters). The names_key file provides a link between the original names and the clean names.
This ensures that the annotations are all the same format, which is useful for the pipelines that reference the annotations (like ProphET).
If working with appropriate organisms, mlplasmids is your best. Otherwise, go with PlasFlow. If you are feeling adventurous, try mob-suite, but use with caution.
Originally I was working on incorporating CAFE, but I later found out that it is not designed to handle certain organisms. So, We went with IslandPath-DIMOB.
See the short version of our head-to-head comparison here: Testing 3 Prophage Finders.
- Q: This seems like a lot of computational time could be saved by simply referencing the coordinates the mobile regions, rather than extracting and running analyses on the subset. A: you're right! but lots of GWAS tools and other rely on fasta input, so it ends up being easier for downstream work. oom for improvment!
- Q: My harddrive is full! why would you do this to me? A: most of these tools rely on their own databases, which are all included in the docker images. Theres no way around it -- thats a loooooot of data.
- Q:
This module was renamed from "mobilephone", it was just too hard to google.
After each run, happie creates a "citing.txt" file, with links to all the software happie uses. Some of those use additional references, so make sure give credit to everyone involved!