MicroPro is a software to perform profiling of both known and unknown microbial organisms for metagenomic dataset. A microbe is considered known/unknown if its whole genome is known/unknown according to the NCBI Reference Sequence Database. MicroPro has two pipelines: MicrobialPip and ViralPip. MicrobialPip considers all the microbial organisms including bacteria, archaea and viruses while ViralPip only extracts known and unknown viruses from the provided metagenomics dataset. More details of MicroPro are provided in the following paper:
- A new release of MicroPro (v1.0.1) is out. It fixes a few bugs in the old version (thanks to JinqunHuang and wangjianshou). Welcome to use!
MicroPro consists of the following four modules:
- Map reads to NCBI Refseq database and perform known microbial species profiling
- Cross-assemble unmapped reads
- Detect virus from assembled contigs (ViralPip only)
- Perform contig binning and unknown organisms profiling
MicrobialPip and ViralPip are mostly similar except that ViralPip has additional procedures (Module 3) for virus detection. In terms of output, MicrobialPip will provide an abundance table for all the microbes while ViralPip only ouputs the abundance table of the viruses.
Depedencies of MicroPro are listed below. You can click the software name to navigate to its website. Note that after installing each dependency, you should add it to path to ensure MicroPro's normal run.
-
Python 3 (>= 3.5) version of Miniconda, Snakemake (== 5.3.0) and python 3 package "psutil" (>= 5.6.3)
-
MicroPro is realized via Snakemake [1], a text-based workflow system. Thus, Snakemake is a must for running any pipeline of MicroPro. With a Python 3 version of Miniconda installed, you can install Snakemake using the following command: (See Snakemake manual for more.)
$ pip3 install snakemake==5.3.0Be sure to install Snakemake with version 5.3.0. A higher version may cause MicroPro to crash.
-
Python 3 package "psutil" is needed when running Snakemake. It can be installed using the following command:
$ pip3 install psutil
-
-
Centrifuge (>= 1.0.3)
-
Centrifuge [2] is used for sequence alignment in Module 1. After installing Centrifuge, you need to download and index the reference genomes for bacteria, archaea and viruses from NCBI Refseq Database with the following commands: (See Centrifuge manual for more.)
$ cd <PATH_TO_CENTRIFUGE> $ centrifuge-download -o taxonomy taxonomy $ centrifuge-download -o library -m -d "archaea,bacteria,viral" refseq > seqid2taxid.map $ cat library/*/*.fna > input-sequences.fna $ centrifuge-build -p 4 --conversion-table seqid2taxid.map --taxonomy-tree taxonomy/nodes.dmp --name-table taxonomy/names.dmp input-sequences.fna abv $ mkdir database $ mv abv* database $ rm -r taxonomy library<PATH_TO_CENTRIFUGE>is the directory where you install Centrifuge. The indexed genomes are located in<PATH_TO_CENTRIFUGE>/databasewith prefixabv. It's recommended not to change the directory name as well as the prefix, since otherwise MicroPro would not locate the reference genomes.
-
-
Megahit (>= 1.1.3)
- Megahit [3] is used for the assembly of unmapped reads in Module 2.
-
R (>= 3.3.1) and three R packages: VirFinder, ShortRead and data.table
- 'VirFinder' [4] utilizes a logistic regression model to differ viral sequences from bacterial sequences.
- 'ShortRead' is a useful R package for input and output of files with various extensions like 'fasta', 'fastq', etc. These file formats are ubiquitous in metagenomics analysis.
- 'data.table' allows fast data.frame operations.
- 'ShortRead' and 'data.table' are required in Module 4.
- 'VirFinder' is required in Module 3 (ViralPip only), meaning that if you just use MicrobialPip, there is no need to install it.
-
BWA (>= 0.7.15) and SAMtools (>= 1.4.1)
- BWA [5] and SAMtools [6] are used for the generation of unknown organisms abundance table in Module 4.
-
MetaBAT 2 (>= 2.12.1)
- MetaBAT 2 [6] is used for contig binning in Module 4.
MicroPro accepts both single-read and paired-end sequenced reads with FASTQ format as inputs. To improve profiling accuracy, it's recommended to perform raw reads preprocessing including adaptor removal and host genome removal.
To install MicrobialPip, clone the github repository and enter MicrobialPip directory with
$ git clone https://github.com/zifanzhu/MicroPro.git
$ cd MicroPro/MicrobialPip
(To install ViralPip, replace the second command with $ cd MicroPro/ViralPip.)
URC, URC-S and FCV are three useful tools required in MicroPro. Their compiled binaries are in folder utils/. To make them executable, run
$ chmod 755 utils/*
In this step, you tell MicroPro the location of your data and other parameters it needs. For single-read data, use
$ R --no-save --file=scripts/parameters.R --args S <PATH_TO_DATA> <FASTQ> <PATH_TO_CENTRIFUGE> <MIN_CONTIG_LENGTH>
For paired-end data, use
$ R --no-save --file=scripts/parameters.R --args P <PATH_TO_DATA> <R1> <R2> <FASTQ> <PATH_TO_CENTRIFUGE> <MIN_CONTIG_LENGTH>
S or P indicates single-read or paired-end data. <PATH_TO_DATA> is the directory of your data. <PATH_TO_CENTRIFUGE> is the directory where you install Centrifuge. It tells MicroPro where to locate the reference genomes. It's recommended to remove the final slash in <PATH_TO_DATA> and <PATH_TO_CENTRIFUGE>. For example, if your data path is '/home/data/'. It's better to replace <PATH_TO_DATA> with /home/data instead of /home/data/. <MIN_CONTIG_LENGTH> refers to the minimum length of the contigs to report in the cross-assembly step. We suggest setting it as 1000. But you may choose whatever fits your study.
<FASTQ> or <R1> <R2> <FASTQ> instructs MicroPro how to generate sample names. <FASTQ> refers to the file extension, like .fastq, .fq, etc. <R1> and <R2> refer to the way that two pairs are represented in the file name, like _1, _2 or _R1,_R2. For single-read data, MicroPro will remove <FASTQ> in the file name and use the rest as the sample name. For paired-end data, MicroPro will separate the file name by <R1> if the file is pair 1 (or <R2> accordingly) and use the part before <R1> (or <R2>) as the sample name. Note that this assumes the parts before <R1> and <R2> are the same for pair 1 and pair 2 file of same sample. If it's not the case, you should rename your data in this way.
For example, for a single-read dataset of 3 samples with file names: A.fastq B.fastq C.fastq, MicroPro will use A B C as three samples' sample names, if you use the following command:
$ R --no-save --file=scripts/parameters.R --args S <PATH_TO_DATA> .fastq <PATH_TO_CENTRIFUGE> <MIN_CONTIG_LENGTH>
For a paired-end dataset of 3 samples with file names: A_1.fq A_2.fq B_1.fq B_2.fq C_1.fq C_2.fq, MicroPro will also use A B C as three samples' sample names, if you use the following command:
$ R --no-save --file=scripts/parameters.R --args P <PATH_TO_DATA> _1 _2 .fq <PATH_TO_CENTRIFUGE> <MIN_CONTIG_LENGTH>
Note that the catenation of strings <sample_name_generated_by_MicroPro>, <R1> (or <R2>) and <FASTQ> must be the same as the corresponding file name, like in previous example, catenating A, _1 and .fq will give you A_1.fq. This means if you have a paired-end dataset of 3 samples with file names: A_1_001.fq A_2_001.fq B_1_001.fq B_2_001.fq C_1_001.fq C_2_001.fq, the following command will result in an error when running MicroPro, since for example the catenation of A, _1 and .fq is not a file name.
$ R --no-save --file=scripts/parameters.R --args P <PATH_TO_DATA> _1 _2 .fq <PATH_TO_CENTRIFUGE> <MIN_CONTIG_LENGTH>
The correct command should be:
$ R --no-save --file=scripts/parameters.R --args P <PATH_TO_DATA> _1_001 _2_001 .fq <PATH_TO_CENTRIFUGE> <MIN_CONTIG_LENGTH>
To test if the installation is complete, run the following codes on a two-sample mock dataset:
$ cd MicroPro/MicrobialPip
$ chmod 755 utils/*
$ R --no-save --file=scripts/parameters.R --args P test/data _1 _2 .fq test 1000
$ snakemake -p -s Snakefile-P
The pipeline takes about 15 min and outputs 6 abundance files under the 'res/' folder if everything is installed properly.
Running MicrobialPip and ViralPip are exactly the same. You just need to make sure that you're in the folder corresponding to the pipeline you want. For single-read data, run
$ snakemake -j <no_cores> -p -s Snakefile-S
For paired-end data, run
$ snakemake -j <no_cores> -p -s Snakefile-P
-j <no_cores> will tell Snakemake to use at most <no_cores> cores. Snakemake supports parallel processing when providing more than one core. This saves lots of time especially when you have a large amount of samples. See Snakemake manual for more Snakemake executing options.
The known and unknown abundance tables are stored in folder res/. Each of them has a 'csv' version as well as a 'rds' version, which can be opened and edited by R. Every table contains a sample-by-organism matrix with each entry representing a known/unknown organism's relative abundance in a sample. For MicrobialPip, 'centrifuge_abundance' is the known abundance table while 'unknown_abundance' is the unknown table. For ViralPip, 'centrifuge_viral_abundance' is for the known while 'unknown_viral_abundance' is for the unknown. Note that a microbe is output in the abundance table only if it appears in at least one sample. Thus, for viral abundance, it's possible that all the viruses of some sample have abundance zero since virus usually makes up a very small proportion of the whole microbial community. Also, combined abundance table (a combination of known and unknown abundances) is output for both versions. It's named 'combined_abundance' for MicrobialPip and 'combined_viral_abundance' for ViralPip.
The intermediate results are stored in corresponding folders. In particular, centrifuge results are stored in 1_centrifuge/; cross-assembly results are in 3_cross_assembly/megahit_out/; VirFinder results (ViralPip only) are in 4_virfinder/4_4_vf_summary/vf_results.rds; Contig binning results are in 5_binning/bins_dir/ for MicrobialPip or viral_3_binning/bins_dir/ for ViralPip.
The intermediate log and benchmark files are stored in logs/ and benchmarks/. The benchmark file records the wall time and the memory usage of a particular intermediate process.
If you use MicroPro, please cite the following paper:
Copyright (C) 2018 University of Southern California
Authors: Zifan Zhu, Jie Ren, Fengzhu Sun
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.
Commercial users should contact Prof. Fengzhu Sun (fsun@usc.edu), copyright at University of Southern California.
[1] Köster, Johannes, and Sven Rahmann. "Snakemake—a scalable bioinformatics workflow engine." Bioinformatics 28.19 (2012): 2520-2522.
[2] Kim, Daehwan, et al. "Centrifuge: rapid and sensitive classification of metagenomic sequences." Genome research (2016).
[3] Li, Dinghua, et al. "MEGAHIT v1. 0: A fast and scalable metagenome assembler driven by advanced methodologies and community practices." Methods 102 (2016): 3-11.
[4] Ren, Jie, et al. "VirFinder: a novel k-mer based tool for identifying viral sequences from assembled metagenomic data." Microbiome 5.1 (2017): 69.
[5] Li, Heng. "Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM." arXiv preprint arXiv:1303.3997 (2013).
[6] Li, Heng, et al. "The sequence alignment/map format and SAMtools." Bioinformatics 25.16 (2009): 2078-2079.
[7] Kang, Dongwan D., et al. "MetaBAT, an efficient tool for accurately reconstructing single genomes from complex microbial communities." PeerJ 3 (2015): e1165.