To separate download (prefetch) and conversion tasks (fasterq-dump).
Example
1) prefetch: download only raw SRA files
Example: download NCBI-SRA sequence data file SRR649944 into SRA_download/ directory
( Older prefetch versions stored SRA files in $HOME/ncbi/public/sra/ )
prefetch SRR649944 -O SRA_download/
ls SRA_download/*/*
SRR649944/SRR649944.sra
2) fasterq-dump: convert SRA to FASTQ
The locally stored SRA files can be converted by fasterq-dump into .fastq files, saved in directory FASTQ_files/
Additionally using gzip to compress .fastq files
fasterq-dump --split-3 SRA_download/SRR649944 -O FASTQ_files/
gzip FASTQ_files/*.fastq
ls FASTQ_files/
SRR649944_1.fastq.gz
SRR649944_2.fastq.gz
# Note, addressing the SRA identifier without exact path may cause re-download
fasterq-dump --split-3 SRR649944
If the SRA file is not located in the current working directory, fasterq-dump will download the file again using the default temporary directory $TMPDIR if no other download location is pre-defined as CACHE in → SRA-toolkit configuration.
ls $TMPDIR/.nfs*
/tmp/.nfs00000000558ac9cf
Options
# set maximum file size download limit to 100G (default: 20G)
prefetch --max-size 100G SRR649944
Read more
https://github.com/ncbi/sra-tools/wiki/HowTo:-fasterq-dump
prefetch --help