G-BLASTN is a GPU-accelerated nucleotide alignment tool based on the widely used NCBI-BLAST.
Main features of G-BLASTN:
- G-BLASTN supports the blastn and megablast modes of NCBI-BLAST, and can produce exactly the same results as NCBI-BLAST.
- Under megablast mode, Nvidia GTX780 can achive an average of 14.8X speedup over single-thread NCBI-BLAST using Intel Core i7-3820 (3.6GHz).
- Under blastn mode, Nvidia GTX780 can achive an average of 4.3X speedup over single-thread NCBI-BLAST using Intel Core i7-3820 (3.6GHz).
- G-BLASTN has very similar user commands as NCBI-BLAST.
- When searching a batch of queries, G-BLASTN supports a pipeline mode that can further shorten your searching time by fully utilizing the GPU and CPU resources.
Limitations of G-BlASTN:
- Currently, The discontiguous megablast mode is not supported.
- The database size should fit into your GPU memory.
- G-BLASTN is not yet suitable for databases with short sequences, such as NCBI nucleotide collection (nt) database.
The interface of G-BLASTN is almost identical to the original NCBI-BLAST with the following additional GPU options:
*** GPU options ***
-use_gpu <true|false>
Use 'true' to enable GPU for blastn
Default = 'false'
-mode <1|2>
1.normal mode, 2.pipeline mode
Default = `1'
-query_list <file>
The file includes the list of file names of your query files, one per line. You can generate the query list file by the following command in the folder that holds your query files:
Linux: find ./ -type f > query_list.txt
Windows: dir /s/b > query_list.txt
Citation:
- K. Zhao and X.-W. Chu, “G-BLASTN: Accelerating Nucleotide Alignment by Graphics Processors,” Bioinformatics, January 2014. [link] [supplementary]
Acknowledgement:
This project is supported by grant FRG2/11-12/158 from Hong Kong Baptist University. We also thank NVIDIA corporation for their donation of GPU cards.