HLA Genotype Imputation with Attribute Bagging

Kernel Version: v1.4

Features

HIBAG is a state of the art software package for imputing HLA types using SNP data, and it relies on a training set of HLA and SNP genotypes. HIBAG can be used by researchers with published parameter estimates instead of requiring access to large training sample datasets. It combines the concepts of attribute bagging, an ensemble classifier method, with haplotype inference for SNPs and HLA types. Attribute bagging is a technique which improves the accuracy and stability of classifier ensembles using bootstrap aggregating and random variable selection.

Bioconductor Package

Release Version: 1.16.0

http://www.bioconductor.org/packages/release/bioc/html/HIBAG.html

Changes in Bioconductor Version (since v1.14.0):

Kernel Version: v1.4
modify the kernel to support the GPU extension
develop a complementary R package (HIBAG.gpu) for GPU computing
The kernel v1.4 outputs exactly the same parameter estimates as v1.3, and the model training with v1.4 is 1.2 times faster than v1.3.

Changes in Bioconductor Version (since v1.3.0):

Kernel Version: v1.3
optimize the calculation of hamming distance using SSE2 and hardware POPCNT instructions if available
hardware POPCNT: 2.4x speedup for large-scale data, compared to the implementation in v1.2.4
SSE2 popcount implementation without hardware POPCNT: 1.5x speedup for large-scale data, compared to the implementation in v1.2.4

Package Author & Maintainer

Dr. Xiuwen Zheng (zhengx@u.washington.edu)

Pre-fit Model Download

http://zhengxwen.github.io/HIBAG/

http://zhengxwen.github.io/HIBAG/hibag_index.html

Platform-specific models: http://zhengxwen.github.io/HIBAG/platforms.html

or http://www.biostat.washington.edu/~bsweir/HIBAG/

Citation

Zheng, X. et al. HIBAG-HLA genotype imputation with attribute bagging. Pharmacogenomics Journal 14, 192-200 (2014). http://dx.doi.org/10.1038/tpj.2013.18

Installation

Bioconductor repository:

source("http://bioconductor.org/biocLite.R")
biocLite("HIBAG")

Development version from Github:

library("devtools")
install_github("zhengxwen/HIBAG")

The install_github() approach requires that you build from source, i.e. make and compilers must be installed on your system -- see the R FAQ for your operating system; you may also need to install dependencies manually.

Install the package from the source code: download the source code

wget --no-check-certificate https://github.com/zhengxwen/HIBAG/tarball/master -O HIBAG_latest.tar.gz
## or ##
curl -L https://github.com/zhengxwen/HIBAG/tarball/master/ -o HIBAG_latest.tar.gz

## Install ##
R CMD INSTALL HIBAG_latest.tar.gz

Acceleration

CPU with Intel Intrinsics

Install the package from the source code with the support of hardware POPCNT (requiring SSE4.2): You have to customize the package compilation, see: CRAN: Customizing-package-compilation

Change ~/.R/Makevars to, if your machine supports SSE4.2 or higher, assuming GNU Compilers (gcc/g++) or Clang compiler (clang++) are installed:

## for C code
CFLAGS=-g -O3 -march=native -mtune=native
## for C++ code
CXXFLAGS=-g -O3 -march=native -mtune=native

Or force to create hardware POPCNT code:

## for C code
CFLAGS=-g -O3 -mpopcnt -msse4.2
## for C++ code
CXXFLAGS=-g -O3 -mpopcnt -msse4.2

If the package compilation succeeds with hardware POPCNT instructions, you should see a welcome message after loading the package:

HIBAG (HLA Genotype Imputation with Attribute Bagging)
Kernel Version: v1.4
Supported by Streaming SIMD Extensions (SSE2 + POPCNT)

GPU with OpenCL

Install HIBAG.gpu from Github:

library("devtools")
install_github("zhengxwen/HIBAG.gpu")

Please use hlaAttrBagging_gpu() and hlaPredict_gpu() for model training and prediction.

Speedup ratios for training HIBAG models:

CPU (1 core)	CPU (1 core, POPCNT)	1x NVIDIA Tesla K80	1x NVIDIA Tesla M60	1x NVIDIA GTX 1080Ti	1x NVIDIA Tesla P100	1x NVIDIA Tesla V100
1	1.63 x	46.5 x	57.5 x	93.7 x	209.1 x	246.3 x

using HIBAG v1.14.0 and HIBAG.gpu v0.9.1

CPU (1 core), the default installation from Bioconductor supporting SIMD SSE2 instructions, using Intel(R) Xeon(R) CPU E5-2630L @2.40GHz

CPU (1 core, POPCNT), optimization with Intel/AMD POPCNT instruction, using Intel(R) Xeon(R) CPU E5-2630L @2.40GHz

The benchmark was made possible, in part, through HPC time donated by Microway, Inc. We gratefully acknowledge Microway for providing access to their GPU-accelerated compute cluster (http://www.microway.com/gpu-test-drive/).

adiamb/HIBAG