LDExplorer is an R package for the memory efficient whole-genome LD-based haplotype block recognition. It is a scalable implementation of the Gabriel et al. 2002 algorithm, which is the most commonly used method for LD-based haplotype block recognition. LDExplorer significantly outperforms the most popular Haploview (Barrett et al., 2005) implementation both in memory usage and runtime. It has only linear memory complexity thus requiring only megabytes of main memory to process genome-wide data (millions of SNPs). LDExplorer omits any restrictions on the maximal haplotype block length and considers LD between SNPs at any distance. The runtime performance was significantly improved by omitting at least 80% of unnecessary computations chromosome-wide. Such an effective search space reduction doesn't alter haplotype block partition and LDExplorer produces results identical to the Haploview implementation.
The following two figures illustrate the efficiency of search space pruning in LDExplorer. Left figure illustrates the complete LD heatmap that is always computed by Haploview. Right figure illustrates only the part of the LD heatmap that is computed by LDExplorer in order to detect haplotype blocks. The haplotype block partitions, indicated with bold lines, are identical between two implementations.
LDExplorer allows to save haplotype blocks in BED format as a genome browser track. Then, the generated BED file can be uploaded into UCSC or Ensembl genome browsers and haplotype blocks will be visualized as depicted in the following figure (UCSC Genome Browser screenshot).
We provide computed genome-wide LD based haplotype blocks for HapMap phase II and 1000 Genomes Project phase I release 3 phased genotypes. The haplotype blocks are saved in BED format and, thus, can be imported into UCSC or Ensembl genome browsers.
Version 1.0.3 - Published 2013-06-11
R version 2.15.1 or higher.
For parallel computations package must be compiled using C++ compiler with OpenMP support.
a.) To install package from CRAN repository, execute the following R command:
b.) To install the binary package on Windows, execute the following R command:
> install.packages("path/to/LDExplorer_X.Y.Z.zip", repos=NULL)
c.) To install the binary package on MacOS X, execute the following R command:
> install.packages("path/to/LDExplorer_X.Y.Z.tgz", repos=NULL)
e.) To install package from source on Linux and MacOS X, execute the following shell command:
$ R CMD INSTALL path/to/LDExplorer_X.Y.Z.tar.gz
The main algorithm is implemented in the mig() function. Additional functions mig_multi_regions() and mig_multi_chr() are for parallel processing of several regions within one chromosome and parallel processing of several chromosomes, respectively.
For an example, execute the following R commands:
The detailed description of the function arguments and output format can be obtained by executing the following R commands:
To transform mig() output into BED file format in order to visualize haplotype blocks in UCSC or Ensembl genome browsers, you can use create_browser_track() function.
The current LDExplorer packages are available here:
Sources and latest development releases are available on project's R-Forge repository.
Previous versions are available on CRAN archive.
Copyright © 2013 by Daniel Taliun, Johann Gamper and Cristian Pattaro. All rights reserved.
LDExplorer 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.
LDExplorer 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 LDExplorer. If not, see http://www.gnu.org/licenses/.
For questions, comments, or any other help regarding the LDExplorer, please contact us through email firstname.lastname@example.org.