CompRanking

CompRanking is a pipeline for comprehensively assessing the AMR key message of metagenomic samples and rank their risk of resistome. CompRanking calculate each AMR key message, antibiotic resistance genes abundance, mobility and potentials to acquired by pathogens at the contigs level. Generally, CompRanking can give results derived from three features as the co-occurrance of ARGs, MGEs on one contigs and their potentials in pathogens.

Getting Started

Installing

Step 1: Change the current working directory to the location where you want the cloned CompRanking directory to be made. Step 2: Clone the repository using git command

git clone https://github.com/GaoyangLuo/CompRanking

Environment settings

Create environment

Please firstly set up all the environment by the following commands. These commands will help to config all the environment needed.

$ cd CompRanking
$ conda env create -f CompRanking.yaml
$ bash setup.sh

Setting conda path

CompRanking relies on multi conda environments. Before run the demo test, conda bin path should be pre-requisit. Please set your absolute bin path of miniconda. For example, your absolute bin path is /home/username/miniconda3/bin.

How to set your absolute conda bin path:

First step, vim your test_yaml.yaml file

$ vi test_yaml.yaml

Second step, Re-write the real path of miniconda/bin

CompRanking:
  abs_path_to_conda_bin: /your_real_path/miniconda/bin #don't use "~" or "./", please use absolute path

Pleast note that don't use relative path, do not use "~" or "./"

Databse download

You can download the databases from the location: https://doi.org/10.5281/zenodo.8073486. Or run the command lines below.

$ wget https://zenodo.org/record/8073486/files/CompRanking_database_v1.tar.gz?download=1
$ wget https://zenodo.org/record/8073486/files/localDB.zip?download=1
$ tar -zxvf CompRanking_database_v1.tar.gz && mv CompRanking_database_v1 databases
$ unzip localDB.zip

Demo test

We provided a set of data for test.

$ python cpr_multiprocess.py -i test_data -t 12 -r 1 -p test_demo

Run gene prediction

Step 1: Gene prediction can generate contextural information of AMR and pathogen information of the whole metagenome. Run the command line below:

$ python cpr_multiprocess.py -i <input_dir> -t <threads> -r <if_restart> -p <project_name_prefix>

Parameters:

-i, <input_dir> contains all the fastq files and fasta files. Files of the the sample should be named using identical <prefix>. For example, FileNameOne_1.fq, FileNameOne_2.fq and FileNameOne.fa represents the pair-end reads fastq files (after quality control) and the assembly file (containing contigs and pleast do not cut into your customed length, default_min_length=500, which cannot be altered).
-t, , the threads you want to use to run the process (Default=16).
-r, <if_restart> 0 or 1. 0 means continue to run after the last break up point. 1 means re-start from the begeining.
-p, <project_name_prefix> You should

Step 2: After finishing all the prediction steps, we should calculate the relative abundance of functional genes, run the command line below:

$ python ./compranking/multiGeneCal_metagenome_rpkg_scg_geneName.py 
        -i <input_dir> 
        -p <project_prefix> 
        -n AGS
        -t 16
        -d <pth2KK2db> #this option is for cell copy normalized by sequence abundance, need to run multiGeneCal_16s.py

The output demo is like below:

ARG_name	Class	Database	MGE_type	Sample_1	Sample_2	Sample_3
AAC(2')-I	aminoglycoside	DeepARG	Unknown	0.003	0.004	0.006
ERMB	macrolide	RGI	phage/plasmid	0.002	0.003	0.004
SUL3	sulfonamide	SARG	plasmid	0.001	0.003	0.005

note: phage/plasmid means ARGs both fpundto be co-located with phage- or plasmid-like contig in one sample (microbial community). plasmid means only found to be co-located with plasmid-like contig. Unknown means not to be found co-located with any MGEs, but not representing it is not co-located with any MGEs, probably due to the accuracy and recall of identification method.

Step 3: Generate a risk score and corresponding valuse of each sample. In this step, you can acquire various parameters such as how many ARGs-carried contigs or phage- or plasmids-related contigs in your samples. Please run the command line below:

$ python ./compranking/baseInfoExtra_nContigs.py -i <input_dir> -p <project_name_prefix>

sample_name/index	nContigs	nARGs_contigs	nMGEs_contig	nMGEs_plasmid_contig	nMGEs_phage_contigs	nPAT_contigs	nARGs_MGEs_contig	nARGs_MGEs_plasmid_contigs	nARGs_MGEs_phage_contigs	nARGs_MGEs_PAT_contigs	fARG	fMGE	fMGE_plasmid	fMGE_phage	fPAT	fARG_MGE	fARG_MGE_plasmid	fARG_MGE_phage	fARG_MGE_PAT	score_pathogenic	score_phage	score_plasmid
Sample2	433650	482	327780	270746	32238	397572	357	304	28	61	0.0011	0.7558	0.6247	0.0746	0.9168	0.0008	0.0007	6.4568	0.0001	23.1492	20.7351	22.7372
Sample1	433650	482	327780	270746	32238	397572	357	304	28	61	0.0011	0.7558	0.6247	0.0746	0.9168	0.0008	0.0007	6.4568	0.0001	23.1492	20.7351	22.7372

How to calculate each ARG class and their carriers counts

Use the jupyter notebook MGE_carried_ARGs_type_count.ipynb to calculate. The metadata record the number of five types of elements that co-exist with ARGs: plasmid, phage, unclassified (can be any type of sequences, including chromosome or other unknown or unidentified MGEs), IS (Insertion Sequence), IE (Integrated Elements). Table will be generated like this:

	Sample1_x	Sample2_y	Sample3_z	Sample3_m	Sample3_n
aminoglycoside	33	35	36	37	38
macrolide	44	45	46	37	38
tetracycline	22	23	24	37	38

sampleName_x: #plasmid
sampleName_y: #phage
sampleName_z: #unclassified
sampleName_m: #IS
sampleName_n: #IE

Re-running if pipeline halted

Every process will generte a checkpointing file in the repo checkdone, with file name like <Your_Project_Name>.index_build.done. If you want to re-run the pipeline from the last broken step, you can set the parameter -r as 0, which means don't re-run from the beginning. If you set 1, means you want to re-run from the beginning. You can also delete the .done file if you want to re-run the speicific step. We make this pipeline able to identify which step you have run and which one is not completed.