RBRA: metagenomic 16S gene assembly

Created by Feng Zeng Summer 2015.

Overview

RBRA is a pipeline for assembling full-length 16S rRNA gene sequences. It is designed for the large-scale metagenomic studies which contain tens to hundreds of samples. RBRA uses phylogenetic tree search and Dirichlet process (DP) mixture clustering techniques to decode the composition of highly similar microorganisms. It is lightweight, ultrafast, accurate and easy-to-use. RBRA works on both Illumina and Roche/454 sequencing data.

Dependency

Following tools are all assumed installed and could be found in system environment:

* Python >= 2.7.10
* GCC >= 4.9.1
* Bowtie2 >= 2.2.4
* Samtools >= 1.2
* Bedtools >= 2.24.0
* Sickle >= 1.33
* Seqtk >= 1.0

Following Python packages are all assumed installed:

* ETE2 Toolkit >= 2.3.1
* Scipy >= 0.15.1
* Numpy >= 1.9.2
* Scikit-bio == 0.4.0
* Biopython >= 1.65

Installation

Set up environment variable:
```
$ export CXX=<PATH TO G++4.9>
```

Clone the repository:

$ git clone https://github.com/homopolymer/RBRA.git

Build and install to $PWD/bin:

$ python setup.py install --prefix=$PWD

16S Database

RBRA depends on the sequences, taxonomic annotations and phylogenetic tree of 16S rRNA genes. GreenGenes (99_otus) contains 203,452 16S rRNA genes, and provides the separate files for sequence, taxonomy and tree. GreenGenes is available at ftp://greengenes.microbio.me/greengenes_release/gg_13_8_otus/. While GreenGenes is constantly used during our development, RBRA works for other annotation databases, e.g. SILVA, if all of sequence, taxonomy and tree files are provided.

To download GreenGenes:

$ bash download_greengenes.sh

Input

The input of RBRA is a metadata file. An example of metadata is given in the <test> directory.

A metadata file, e.g. data_info.txt, contains the following contents:

Sequencing data
1. BamFiles=<FILE>, a file listing the mapping files of data against 16S rRNA reference, one BAM file per line.
```
$ ls -A1 *_to_gg_99_otus.bam | xargs realpath > bams.fofn
```
Gene data
1. GeneSeq=<FILE>, a FASTA file storing the reference sequences of 16S rRNA genes.
2. GeneIndex=<FILE>, the FASTA index file.
3. GeneTree=<FILE>, a Newick file recording the phylgenetic tree of 16S rRNA genes.
4. GeneTax=<FILE>, a taxonomic annotation of 16S rRNA genes.
5. GeneAlign=<FILE>, a FASTA file of 16S sequence alignments.

Usage

$ python rbra.py [-c cores] [-v] [-p output_prefix] data_info.txt

Test

The repository RBRA_Test contains data and script to test whether RBRA installs and works correctly. Download test data and run the following codes for test:

$ cd test
$ python make_datainfo.py
$ python <PATH TO rbra.py> -c 20 -v data_info.txt 2>&1 | tee log.txt
$ less log.txt

Change Log

September 7, 2015, RBRA v0.3.1 released
1. Add a filtration to StrainCall to filter out noisy assemblies
September 7, 2015, RBRA v0.3.0 released
1. Add coverage fraction into seed gene reference finding
2. Add a script to count 16S assembly abundance per sample
September 2, 2015, RBRA v0.2.0 released
1. Replace mpi4py by multiprocessing
2. Change mawk to awk
3. Clean code
August 31, 2015, RBRA v0.1.0 released

Development Team

Feng Zeng, Xiamen University
Zicheng Wang, Tsinghua University
Ting Chen, Tsinghua University

homopolymer / RBRA