CoLoRMap: Correcting Long noisy Reads by Mapping short reads

Installation

In order to install CoLoRMap, you should first fetch the source code from CoLoRMap git repository.

git clone --recursive https://github.com/sfu-compbio/colormap.git

Please note that command argument --recursive is necessary for downloading submodules automatically. After obtaining the code, you need to install the dependencies. CoLoRMap uses BWA, SAMtools, and Minia. In order to build these dependencies, change to the source directory colormap and use make deps command.

cd colormap
make deps

At last, you can compile CoLoRMap binaries simply by running make command.

make

CoLoRMap corrects long reads in two different steps: (i) using a shortest path (SP) algorithm. (ii) using an One-End Anchor (OEA) based algorithm.

Preparing the short read data

SP algorithm does not need paired-end information, but OEA algorithm actually uses paired-end information. In both cases, the program expects to be fed with a single short read file. In case of OEA algorithm, program expects paired-end short reads in interleaved/interlaced format.

Usually, paired-end short reads are stored in two different files. A single interleaved/interlaced read file can be obtained using fastUtils program which can be found in bin directory after building the program:

cd testData
../bin/fastUtils shuffle -1 ill_1.fastq -2 ill_2.fastq -o ill.fastq

Correcting long reads

To correct long reads, you can use runCorr.sh script:

../runCorr.sh pac.fasta ill.fastq testCorr pre 4

This runs shortest path correction algorithm for long reads stored in pac.fasta by short reads stored in ill.fastq using 4 threads. When this is done, the corrected long reads are stored in testCorr/pre_sp.fasta file.

Improving the correction using One-End Anchors (OEAs)

The script runOEA.sh can be used to further improve the quality of corrected long reads by using One-End Anchors (OEAs) to extend the borders of the corrected regions.

../runOEA.sh testCorr/pre_sp.fasta ill.fastq testOEA pre 4

This runs OEA algorithm for pre-corrected long reads stored in testCorr/pre_sp.fasta by paired-end short reads stored in interleaved/interlaced format in ill.fastq using 4 threads. When this is done, the corrected long reads are stored in testOEA/pre_oea.fasta file.

Publication

Haghshenas E., Hach F., Sahinalp S.C. and Chauve C., "CoLoRMap: Correcting Long Reads by Mapping short reads" Bioinformatics (2016) 32 (17): i545-i551 DOI: 10.1093/bioinformatics/btw463

Contact

Please report problems and bugs on issues page. Otherwise, contact ehaghshe[at]sfu[dot]ca