Menu ▾ ▴

The OPERA wiki

Gao Song Niranjan Nagarajan Denis Bertrand Koh Jia Yu

Welcome to the OPERA wiki!

Introduction

OPERA (Optimal Paired-End Read Assembler) is a sequence assembly program (http://en.wikipedia.org/wiki/Sequence_assembly). It uses information from paired-end/mate-pair/long reads to order and orient the intermediate contigs/scaffolds assembled in a genome assembly project, in a process known as Scaffolding. OPERA is based on an exact algorithm that is guaranteed to minimize the discordance of scaffolds with the information provided by the paired-end/mate-pair/long reads (for further details see Gao et al, 2011).

Note that since the original publication, we have made significant changes to OPERA (v1.0 onwards) including refinements to its basic algorithm (to reduce local errors, improve efficiency etc.) and incorporated features that are important for scaffolding large genomes (multi-library support, better repeat-handling etc.), in addition to other scalability and usability improvements (bam and gzip support, smaller memory footprint). We therefore encourage you to download and use our latest version: OPERA-LG. In our benchmarks, it has significantly improved corrected N50 and reduced the number of scaffolding errors. Furthermore, our latest release contains the wrapper script OPERA-long-read that enables scaffolding with long-reads from third-generation sequencing technologies (PacBio or Oxford Nanopore). The manuscript describing the new features and algorithms is available at Genome Biology. We look forward to getting your feedback to improve it further.

Updates

Release 2.0.6 (18-Nov-2016)

Minor change: Samtools folder can be specified on OPERA-LG.

The latest release can be downloaded here.

Release 2.0.5 (26-Aug-2016)

Updated version of the wrapper script OPERA-long-read.pl
1) Repeat detection module based on short and long-reads;
2) Scaffold using contig links derived from both short-read and long-read libraries.

Updated version of the wrapper script preprocess_reads.pl
1) Changes in the passage of parameters;
2) Allows users to provide the path to samtools executables.

It can be downloaded here.

Dependencies

  1. samtools (only version 0.1.19 or below)
  2. bwa
  3. blasr (only version 1.3.1 or below)

Installation

Type "make install" in the root directory of OPERA-LG.

Typical Usage

Input

  1. Assembled contigs/scaffolds in multi-fasta format (e.g. test_dataset/contigs.fa). Note that for Velvet and SOAPdenovo output, OPERA-LG automatically recognizes repeat contigs and filters them out. For other assemblers, the ideal input is a set of non-repeat contigs though OPERA-LG can use coverage information from mapped reads to identify potential repeat contigs.
  2. Paired-end reads to be used for scaffolding (e.g. test_dataset/lib_1_1.fa, test_dataset/lib_1_2.fa).

Running OPERA-LG

1) Reads need to be mapped onto contigs (currently we provide a script that uses bowtie or bwa):

    perl bin/preprocess_reads.pl 
    --contig <contig-file> 
    --illumina-read1 <read-file-1> 
    --illumina-read2 <read-file-2> 
    --out <output-file> 
    --map-tool <mapping-tool>

where read-file-1 and read-file-2 contain paired-end reads in fasta or fastq format. Mapping-tool should be either bwa (default) or bowtie.

The wrapper assumes that bwa or bowtie, and samtools binaries are found in your PATH. Otherwise, you may specify the location to the binaries by adding the respective arguments: --tool-dir --samtools-dir

For example:

perl bin/preprocess_reads.pl 
  --contig test_dataset/contigs.fa 
  --illumina-read1 test_dataset/lib_1_1.fa 
  --illumina-read2 test_dataset/lib_1_2.fa 
  --out test_dataset/lib_1.map

2) There are two ways to provide parameters to OPERA-LG:

a. Using the command line

           bin/OPERA-LG <contig-file> <mapping-files> <output-folder>

           <contig-file>           Multi-fasta contig/scaffold file
           <mapping-files>         Comma-separated list of files containing mapping of 
                                   paired-end reads for each mate-pair/paired-end library
           <output-folder>         Folder to save scaffold results in
           <samtools-dir>          Folder which contains samtools binaries 
                                   (assumed on PATH if unspecified)

For example:

           bin/OPERA-LG test_dataset/contigs.fa test_dataset/lib_1.map,test_dataset/lib_2.map 
           test_dataset/results

b. Using a configuration file 

           bin/OPERA-LG <config-file>

           <config-file>           Configuration file

For example:

           bin/OPERA-LG test_dataset/multiLib.config

where the configuration file provides information on the contig file, mapping files and output directory to use (in addition to other optional parameters; see below for the format).

Running OPERA-LG for long reads:

The wrapper script "OPERA-long-read.pl" enables OPERA-LG to scaffold contigs using short paired-end reads and long-reads from third-generation sequencing technologies. The mapping of long reads (PacBio or Oxford Nanopore) is performed using blasr. The contig links are then derived from the long-read mapping using the approach described in the Supplementary Note 2 of the OPERA-LG paper. The latest version include a 2 steps repeat detection module: (1) short-reads are mapped to the assembly and repeat contigs are flagged according to their deviation to the average assembly coverage, (2) using long-reads mapping, contigs with conflicting adjacent neighbors are detected and flagged as repeats.

The wrapper can be called using the following command line:

 perl bin/OPERA-long-read.pl 

    --contig-file <fasta file of contigs> --kmer <value of kmer used to produce the contigs> 
    --illumina-read1 <fasta file of Illumina read 1> 
    --illumina-read2 <fasta file of Illumina read 2>
    --long-read-file <fasta file of long reads> 
    --num-of-processors <number of processors for the mapping steps>
    --output-prefix <prefix of output mapping file> 
    --output-directory <output directory for scaffolding results>

For example:

 perl bin/OPERA-long-read.pl 

    --contig-file test_dataset_long_reads/contigs.fa 
    --illumina-read1 test_dataset_long_reads/illumina_1.fastq.gz 
    --illumina-read2 test_dataset_long_reads/illumina_2.fastq.gz
    --long-read-file test_dataset_long_reads/nanopore.fa 
    --output-prefix opera-lr --output-directory RESULTS

The wrapper assumes that bwa or bowtie (one of this tool is required for the mapping of the short-read library), blasr (required for the mapping of the long-read library), samtools and OPERA-LG binaries are found in your PATH. Otherwise, you may specify the location to the binaries by adding the following arguments to OPERA-long-read.pl : --short-read-tooldir --blasr --samtools-dir --opera. For the short read mapping bwa is taken as the default mapper. The mapping tool can be specified using: --short_read_maptool.

Output Format

Scaffolds output by OPERA-LG can be found in a multi-fasta file "scaffoldSeq.fasta". Summary assembly statistics can be found in the file "statistics".

Format of Configuration File

An example configuration file can be found in "multiLib_example.config". The main parameters that need to be specified are:

a) contig_file: a multi-fasta file containing assembled contigs/scaffolds (input to OPERA-LG).
b) map_file: the mapping file specifying the location of paired-end reads on the contigs/scaffolds
(input to OPERA-LG; see bin/preprocess_reads.pl).
c) output_folder: the directory into which all results are written.
d) samtools_dir: folder which contains samtools binaries (assumed on PATH if unspecified)
e) kmer: the value of kmer used to produce the assembled contigs/scaffolds (input to OPERA-LG). If not specified, OPERA-LG will try to analyze corresponding assembly file (LastGraph for Velvet assembly or <prefix>.preGraphBasic for SOAPdenovo assembly) in the same directory containing the contig_file. Kmer will be set to 100 if the file cannot be found.</prefix>

Test Datasets

We provide 2 test datasets:

  1. test_dataset.tar.gz that contains paired-end and mate-pairs libraries that can be processed using OPERA-LG.
  2. test_dataset_long_reads.tar.gz that contains paired-end and Oxford Nanopore reads that can be processed using OPERA-long-reads.pl.

Related Tools

=========

  • FinIS: FinIS is a tool for gap-closing and in silico assembly validation. It can be downloaded here.

  • Sigma/OperaMS: OperaMS is an extension of Opera for Metagenomic Assembly that is based on Sigma, an algorithm for clustering metagenomic assemblies based on multiple sources of information. It can be downloaded here.

References

  • To cite OPERA please use the following citations:

Song Gao, Denis Bertrand, Burton K. H. Chia and Niranjan Nagarajan. OPERA-LG: efficient and exact scaffolding of large, repeat-rich eukaryotic genomes with performance guarantees. Genome Biology, May 2016, doi: 10.1186/s13059-016-0951-y.

Song Gao, Wing-Kin Sung, Niranjan Nagarajan. Opera: reconstructing optimal genomic scaffolds with high-throughput paired-end sequences. Journal of Computational Biology, Sept. 2011, doi:10.1089/cmb.2011.0170.

Please feel free to contact us if you find bugs, have suggestions, need help etc. Use the discussion forum, the mailing-list or simply mail us directly.

Note on older releases

======

Release 2.0.4 (26-May-2016)

Changes from version 2.0.2:

  • The new wraper script OPERA-long-read.pl for scaffolding with long reads from third-generation sequencing technologies (PacBio or Oxford Nanopre).

It can be downloaded here.

Release 2.0.2 (21-Mar-2015)

Changes from version 2.0.1:
* Format of mapping files will be checked before analysing. An error will be reported if the format
(e.g. column number) is not corerect;
* In preprocess_reads.pl, users can specify the temporary directory for sorting mapping files (default
directory is current directory).

It can be downloaded here.

Release 2.0.1 (21-Mar-2015)

Changes from version 2.0:

  • Polyploid sequence assembly is supported by specifying "ploidy" value in a configuration file (the coverage of haploid sequence can also be specified through "haploid_coverage", but preferred to be calculated by OPERA-LG);
  • The minimum scaffolded contig length is max( 500, 2 * smallest pair-end library mean);
  • Multiple libraries will be used in three groups: smaller than 1kbp, between 1kbp and 10kbp, bigger than 10kbp.

It can be downloaded here.

Release 2.0 (2-Jul-2014)

Changes from version 1.4:

  • Repetitive contigs can be scaffolded through adding "filter_repeat=no" into configuration file;
  • Distances of edges are re-estimated based on the fact that the observed read pairs are from a truncated distribution.

It can be downloaded here.

Release 1.4 (19-Sep-2013)

Changes from version 1.3.1:

  • Search space in Opera is better prioritized to significantly reduce the number of local scaffolding errors;
  • Multiple libraries are used simultaneously to reduce scaffolding errors further;
  • Compressed read files (suffix is ".gz") can be handled by preprocess_reads.pl;
  • Read mappings are now stored in the "bam" format;
  • Opera supports the direct reading of read mappings in the "bam" format.

It can be downloaded here.

Release 1.3.1 (15-Jan-2013)

Changes from version 1.3:

  • Fixed a timing out bug in Opera.

It can be downloaded here.

Release 1.3 (20-Dec-2012)

Changes from version 1.2:

  • Improved library mean and standard deviation calculation.

It can be downloaded here.