Add snippet for fastq-dump.

protocols-computer/snippets/random-computer-snippets.rst

@@ -4,6 +4,8 @@
 Random Computer Snippets
 ========================
 
+All of the following assume that you are using the Z shell (zsh).  These may or may not work in BASH.
+
 Subsample reads for R1 and R2 using seqtk
 -----------------------------------------
 
@@ -19,4 +21,56 @@ Subsample reads for R1 and R2 using seqtk
 			echo $file;
 			seqtk sample -s $RAND $file $READS | gzip > $file:t:r:r.SUBSAMPLE.fastq.gz
 		done;
-	done
+	done
+
+Download data for multiple files from NCBI SRA
+----------------------------------------------
+
+First, create a list of SRRs in a file, `sra-records.txt`, that looks something like:
+
+.. code-block:: text
+
+	SRR453553
+	SRR453556
+	SRR453559
+	SRR453277
+	SRR453409
+	SRR453550
+	SRR452995
+	SRR453269
+	SRR453270
+	SRR453274
+	SRR453263
+
+`fastq-dump` is slow AF over the internet. It's faster to download the SRA file using wget, then parse that into fastq:
+
+.. code-block:: bash
+
+	for record in `cat sra-records.txt`; 
+	do
+		echo $record;
+		wget ftp://ftp-trace.ncbi.nih.gov/sra/sra-instant/reads/ByRun/sra/SRR/${record[1,6]}/$record/$record.sra;
+		fastq-dump -I --split-files --gzip $record.sra;
+		rm $record.sra;
+	done
+
+This is still pretty darn slow.  We can parallelize using GNU `parallel`.  After creating `sra-records.txt` to hold the list of files that we need, create a script named `get_sra.sh` with the following contents:
+
+.. code-block:: bash
+
+	#!/bin/zsh
+
+	echo "starting $1";
+	wget ftp://ftp-trace.ncbi.nih.gov/sra/sra-instant/reads/ByRun/sra/SRR/${1[1,6]}/$1/$1.sra;
+	fastq-dump -I --split-files --gzip $1.sra;
+	rm $1.sra;
+	echo "finished $1";
+
+Once, that's completed, make it executable (`chmod 0755 get_sra.sh`). Now, you can run this on 12 cores (`--jobs 12`) using:
+
+.. code-block:: bash
+
+	parallel --progress --joblog logfile --jobs 12 -a sra-records.txt ./get_sra.sh {$1}
+
+This will download the files from SRA and convert them to fastq.gz in parallel.
+