MIG-finder

Metagenomic Integron-associated Gene finder

Introduction

MIG-finder is a tool developed to parse genomic data searching for integron-associated genes. The method is optimised for partly assembled metagenomic data but can be used in whole genomes as well. Integrons are genomic mobile elements carry an integrase gene, a promoter and an array of gene cassettes. Because of the fragmented nature of (assembled) metagenomic data, MIG-finder do not look for the whole integron, instead it searches for their gene cassettes and require at least two to be found in the same sequence so that the false positive number is reduced.

Method

MIG-finder is thus a bioinformatics pipeline. It starts by searching for attC sites using HattCI, validates these sites checking their secondary structures using Infernal, filters out single hits to reduce number of false positives and finally search for ORFs on the vicinity of the attC sites using Prodigal.

Requirements

MIG-finder is a python script and requires other tools to be installed and available to run properly. You will need:

• HattCI 1.0b+
• Infernal 1.1.1+
• Prodigal 2.6.2+

Instalation

Download and uncompress the files provided here. In the directory where migfinder directory is located run:

user@machine:~$ pip install -e migfinder

Alternatively, clone the repository and then:

1. unzip the folder
2. cd migfinder-master/migfinder
3. run make

Running

Recommended to create a folder to each fastafile analysed. From this folder, start python and run:

import migfinder as mf
mf.main("/path/to/fastafile")

Note that you need to provide the full path to your fasta even if you have it in your cwd.

You can opt to give MIG-finder plenty of arguments:

• fastafile an input .fasta, .fna, or .fa containing the genomic sequences that may contain integron-associated genes. Note that the fasta file needs to be located inside migfinder-master/migfinder
• both logical value that indicates if both the ordinary sequences and the complementary sequence should be processed [default: True]
• nseq HattCI reads x sequences at a time and processes them before reading the next x sequences, in order to avoid overextending RAM. This flag gives the option to manually choose number of sequences to read, in the case of large sequences, reduce it [default: 1000]
• nthread HattCI may run a large part of the computations in parallel, i.e. let different threads process a set of sequences, which in turn gives a reduced computation time. The parallelization works best when processing larger chunks of sequences at a time [default: 6]
• cm_model covariance model used by Infernal to validate the attC site secondary structure. A model is provided with the implementation and it is recommended to be used unless you have a good reason use an alternative one.
• k_cm threshold used to filter Infernal results [default: 20]
• k_orf threshold used to filter HattCI results [default: 0]
• save_orf save ORF results in a separate fasta file [default: True]
• dist_threshold max distance allowed to consider two adjacent attC sites part of the same integron [default: 4000]
• d_CDS_attC max distance allowed between ORF and attC site in the same gene cassette. Used to validate the first attC site in the array [default: 500]

Example data

Use the file mgm4570162.3.050.1.fasta.zip provided in the package to test your installation. Notice you need to unzip the file first :)

Output

The output is organised in:

• hmmresults, dir with HattCI results
• cmrsults, dir with Infernal results
• orfresults, dir with Prodigal results, only if attC sites have been found.

If attC sites have been found, the final results will be organised in the fastaname.results file. The columns in this file are:

id	sequence id from fastafile
element	if the hit is a CDS (ORF) or attC site
start	start position of the element in the original fastafile
end	end position of the element in the original fastafile
strand	strand of the element in the original fastafile
score	Prodigal score for ORFs, Infernal score for attC sites
len	length of the element
dist	distance to previous element. ini0 indicates first in the array
array_no	CDS_# or attC_#, where # is the number of the gene cassette in the array they belong to
dist_attC	distance between attC sites, from the end of one to the start of another
Vscore	Viterbi score given by HattCI
R', sp', L', loop, L", sp", R"	start position of each one of the attC site motifs