************************ G-BLASTN 1.1 [November 2013] ****************************
G-BLASTN is a GPU-accelerated nucleotide alignment tool based on the widely used NCBI-BLAST.
G-BLASTN can produce exactly the same results as NCBI-BLAST, and it also has very similar user
commands. It also supports a pipeline mode, which can fully utilize the GPU and CPU resources
when handling a batch of queries. G-BLASTN supports megablast and blastn mode. The discontiguous
megablast mode is currently not supported.
G-BLASTN is free software and you can browse/download the source code at:
https://sourceforge.net/p/gblastn
or
https://github.com/OpenHero/gblastn
News:
1.1 version release is built on NCBI-BLAST 2.2.28.
I. Supported features
=====================
G-BLASTN accelerates the blastn module of NCBI-BLAST by GPUs. G-BLASTN has been tested on
NVIDIA GPUs GTX680, GTX780, and Quadro K5000.
Requirement:
============
1). Nvidia GPU card with compute capability > 1.3
2). CUDA5.5 version. https://developer.nvidia.com/cuda-toolkit
II. Installation instructions
=============================
G-BLASTN directly modifies NCBI-BLAST by adding GPU functionality. To install G-BLASTN, you can:
1) Download v1.1.tar.gz of gblastn from https://github.com/OpenHero/gblastn/archive/v1.1.tar.gz
and unpack the package:
For example, on 64-bit Linux:
>tar zxvf v1.1.tar.gz
>cd gblastn.1.1
>chmod +x install
2) Install G-BLASTN
>./install
This will:
i. Ask the user whether G-LBASTN should be added to an existing BLAST installation or
whether NCBI-BLAST should be installed as well.
ii. Modify the existing NCBI-BLAST installation or download, unpack and unzip NCBI-BLAST from
ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/2.2.28/${ncbi_blast_version}.tar.gz
depending on what was selected by the user in (i).
iii. Compile the CUDA files.
iv. Embed G-BLASTN into the existing or downloaded NCBI-BLAST.
If the installation is successful, you should find the executable "blastn" in
"./ncbi-blast-2.2.28+-src/c++/GCC447-ReleaseMT64/bin/".
NOTE: The directory "GCC447-ReleaseMT64" might differ on your system.
Acknowledgement: The installation configuration of G-BLASTN is based on that of GPU-BLAST
(http://eudoxus.cheme.cmu.edu/gpublast/gpublast.html).
3) The G-BLASTN
If there is no error, you can get the binary G-BLASTN file "blastn" in directory "/ncbi-blast-2.2.28+-src/c++/GCC447-ReleaseMT64/bin/".
Then move the "blastn" file into "bin" directy by command "mv" as follows:
First go to the "/ncbi-blast-2.2.28+-src/c++/GCC447-ReleaseMT64/bin/" directory, then type
>mv blastn /home/blsatn/bin/gblastn
III. How to use G-BLASTN
========================
The interface of G-BLASTN is almost identical to the original NCBI-BLAST with the following additional options:
*** GPU options
-use_gpu <true|false>
Use 'true' to enable GPU for gblastn
Default = 'false'
-mode <1|2>
1.normal mode, 2.pipeline mode
Default = `1'
-query_list <file>
The file includes the list of files names of your query files.
IV. Example:
============
1) In the work directory, we can make some sub-directories:
.
├── bin
│ ├── blastn # the orignal blastn
│ └── gblastn # the G-BLASTN (you need to copy the compiled binary "blastn" into this direcotry, and rename to "gblastn" or others;
├── blast
│ └── src
│ └── gpu
│ ├── c++
│ ├── install
│ ├── log
│ ├── ncbi-blast-2.2.28+-src
│ └── 2.28.zip
├── data # NCBI nt database
│ └── nt2m
│ ├── nohup.out
│ ├── nt.2m.00.nhr
│ ├── ...
│ ├── nt.2m.counts
│ └── nt.2m.nal
├── database # human and mouse database
│ └── blastdb
│ ├── human
│ │ ├── human.1-8.wm.body
│ │ ├── ...
│ │ ├── human.9-Y.wm.nsq
│ │ └── human.wm.nal
│ └── mouse
│ ├── mouse.1-10.wm.nhr
│ ├── ...
│ ├── mouse.11-Y.wm.nsq
│ ├── mouse.nal
│ └── mouse.wm.nal
├── output # output directory
│ ├── query.101.fa.out
│ └── ...
├── query # query directory
│ ├── queries
│ │ ├── human
│ │ │ ├── qlarge
│ │ │ │ ├── query.11.fa
│ │ │ │ └── ...
│ │ │ ├── qmedium
│ │ │ │ ├── query.100.fa
│ │ │ │ └── ...
│ │ │ └── qsmall
│ │ │ ├── query.101.fa
│ │ │ └── ...
│ │ └── mouse
│ │ ├── qlarge
│ │ │ ├── query.201.fa
│ │ │ └── ...
│ │ ├── qmedium
│ │ │ ├── query.101.fa
│ │ │ └── ...
│ │ └── qsmall
│ │ ├── query.100.fa
│ │ └── ...
│ └── SRR955707
│ ├── 1.fa
│ └── ...
├── script # shell directory
│ └── new_query_human_mouse
│ ├── g.m.sh
│ └── m.ls
└── wm_counts # mask database for human and mouse
├── human.stat
└── mouse.stat
2) The shell of g.m.sh
###################################################################################################################
#!/bin/sh
echo "***************start*****************"
blastn="/home/blastn/bin/gblastn"
data_dir="/home/blastn/"
database="${data_dir}database/blastdb/mouse/mouse.wm"
maskdb="${data_dir}wm_counts/mouse.stat"
query_list="m.ls"
outpath="/home/blastn/output/log.txt"
$blastn -db $database -window_masker_db $maskdb -query_list $query_list -outfmt "7" -out $outpath -use_gpu true
echo "***************complete*****************"
###################################################################################################################
3) The m.ls file contains the path names of all your queries, e.g.:
/home/blastn/query/queries/mouse/qmedium/query.195.fa
/home/blastn/query/queries/mouse/qmedium/query.130.fa
