How can I find SRA MD5 checksums for FASTQ files?

Question

[rivanov]

I've been searching for some time, but I cannot find any md5sums that come with FASTQ files from the SRA. I'm currently using sra-tools' fasq-dump to get them from a script. On their website, only analyses seem to have md5sums.

[Ram]

OP has not been seen in 16 months. I request a mod to review the answers and pick accepted answers among them.

[Ram]

Thank you, GenoMax!

[osulliva]

There are many mays to customize fastq output from SRA. So there is no single checksum that would be relevant

Answer 1

[Matt Shirley]

The SRA archive format ("vdb") contains an md5 checksum as well as a few other consistency checks (I think). The sra-toolkit has a utility, vdb-validate which will report any errors in the data, and perform an md5 checksum comparison.

Answer 2

[Shicheng Guo]

Try this one. When you didn't remove SRA files, you can use the following:

for i in `ls *.gz`; 
do 
  SRR=${i%%_*}; 
  echo $SRR >> list.txt; 
done

for j in `sort -u list.txt`
do
  vdb-validate $j
done

if you have already removed SRA and only kept fastq, and suppose your fastq is from this way:

fastq-dump --split-files --gzip SRR949210

Then you can try:

for i in `ls *fastq.gz`; 
do 
gunzip -t $i 2 > $i.err
done
find . -name "*err" -type f -size +0c -exec -larth {} \;

Answer 3

[osullivanchristopher]

vdb-validate operates on the SRA archive. if you are streaming fastq from SRA using local toolkit install, you'll get an error if the process is interrupted. otherwise your output will be complete. you can verify by checking your output stats against the stats stored in SRA using sra-stat -x (use --statistics to get deeper detail including at readgroup level and quality score distribution). or check individually by looking at SRA RunBrowser which displays basic stats.