DefenseFinder is a program to systematically detect known anti-phage systems. DefenseFinder uses MacSyFinder.
If you are using DefenseFinder please cite
This repository contains DefenseFinder a tool allowing for a systematic search of anti-phage systems. The DefenseFinder models based on MacSyFinder architecture can be here
DefenseFinder has one program dependency: the Hmmer program, version 3.1 or greater (http://hmmer.org/). The hmmsearch program should be installed (e.g., in the PATH) to use MacSyFinder. DefenseFinder also relies on Python library dependencies:
DefenseFinder is installable through pip. Before starting, if you can, it is recommended to install DefenseFinder in a virtualenv (such as condas).
conda create –name defensefinder
conda activate defensefinder
pip install mdmparis-defense-finder
However, you can also install DefenseFinder using only pip.
pip install mdmparis-defense-finder
At this stage, if you have an issue, this could be due to a problem with your pip installer. Check the following webpage for details on how to solve it
After installing DefenseFinder, you need to retrieve the DefenseFinder models. To retrieve it run:
defense-finder update
When you have not used DefenseFinder in the last days, make sure you have the latest versions of the models. To verify and downloaded if necessary the latest models run:
defense-finder update
When the DefenseFinder models are updated you only have to update the models and not the tool. However, if you have an outdated version of the DefenseFinder tool, you can use the following line to get the most recent version
pip install -U mdmparis-defense-finder
defense-finder update
If you want to run DefenseFinder on a small set of genomes (< 30 000 proteins). You can run the following command.
defense-finder run genome.faa
The input file, here “genome.faa” has to be under the format of protein fasta, where all proteins are in the order of their position in the genome. Indeed DefenseFinder takes into account the order of the proteins.
A run on a genome (few thousand proteins) should take less than two minutes on a standard laptop. If more, make sure everything is installed properly. In this configuration, all the replicon will be named UserReplicon. ATTENTION, If you want to run DefenseFinder on a larger set of genomes you need to format your dataset as described in “Larger dataset and Gembase Format”.
DefenseFinder will generate three types of files (and an option to conserve MacSyFinder options). All the files are described below.
defense_finder_systems.tsv : In this file, each line corresponds to a system found in the given genomes. This is a summary of what was found and gives the following information
defense_finder_genes.tsv : In this file, each line corresponds to a gene found in a system. For each gene, there is several information such as the replicon, the position, the system.. All the information comes from MacSyFinder and follows MacSyFinder nomenclature (best_solution.tsv) and more can be found in the MacSyFinder Ma documentation.
defense_finder_hmmer.tsv : In this file, each line corresponds to an HMM hit. This file show all hit of HMM regardless if they are in a complete system or not. Those results have to be used cautiously for deep inspection. Indeed, biologically, it was shown that only a full system will be anti phage. This function can be used to found defense gene in small portion of genomes. Beware, a single protein can have several hits. The output is a part of the result of HMMer results table. - hit_id : the protein name (name found in the input file) - replicon : The name of the replicon - position_hit: The position in the input file - Gene_name : the name of the HMM
By using the argument –preserve-raw , you will have all the results from MacSyFinder. Those results are explained here
When running DefenseFInder on several genomes, like MacSyFinder, we propose to adopt the following convention to fulfill the requirements for the “gembase db_type”. The difference between any fasta file and the gembase format is the name of the protein (Protein name = the text after > in a fasta file). For both type, protein have to be ordered but in the first case the name of the protein do not matter (except no repetition). In the gembase format, protein name are composeded of two part: the replicon and the position. The replicon name is the same for all the protein that the user want to be analyse simultaneously (for example a complete genome, a plasmid…) The second component is the position. Those two component must be separated by “". It is possible to use "” in the replicon name, only the last instance will be used as the separator between replicon name and position. With this format of file, MacSyFinder will be able to treat each replicon separately to assess macromolecular systems presence and reduce ressource use.
Example: esco_genomes.faa
> ESCO388_0001
XXXXXXX
> ESCO388_0002
XXXXXXX
..
…> ESCO388_3603
XXXXXXX
> ESCO389_0001
XXXXXXX
> ESCO389_0002
XXXXXXX
..
…> ESCO389_3555
XXXXXXX
To use DefenseFinder with gembase format file on larger dataset of genomes run
defense-finder run –dbtype gembase esco_genomes.faa
To check the different DefenseFinder options run
defense-finder run --help
DefenseFinder has 4 options:
For questions: you can contact aude.bernheim@inserm.fr