This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Analyze RNA-data directly from SRA database and through Galaxy web tool

Hi all,

I am new to RNA-seq & I have a narrow-band (256 kb) internet access. Is there a way to analyze RNA-seq data (from SRA) without downloading them? Especially through Galaxy (https://usegalaxy.org/)?

Thank you indeed very much.

rna-seq galaxy

2 answers

Galaxy questions are better answered on the Galaxy site. Having said that, there's a tool under "Get Data" on the left called "EBI SRA". That can be used to load files from ENA, which largely has the same files as SRA, into Galaxy without using your bandwidth (at least it shouldn't). I imagine that this is throttled, so it still may take a while to get things.

Thank you indeed very much. That works and I did not need to download anything. Perfect.

Here's an example of how to analyze RNA-seq data from SRA (without using Galaxy and without downloading the data!). First you need four things.

  1. Obtain an SRR accession of interest, i.e. a set of RNA-seq runs from the SRA database at NCBI. Try browsing the SRA website, then use the "Send to" pulldown and send the results to the Run Selector to help choose SRR accessions from your favorite data set(s).
  2. Obtain HISAT, a really fast aligner.
  3. The HISAT installation instructions tell you that to use SRA data directly you also need to install NCBI-NGS toolkit, and you should follow the installation details given in the HISAT manual.
  4. Obtain a reference genome (e.g. I use chromosome 20 because it's conveniently small, in this case below from Ensembl, but you can try it from NCBI or UCSC as well).

Prepare your reference genome:

$ hisat-build Homo_sapiens.GRCh38.dna_rm.chromosome.20.fa Homo_sapiens.GRCh38.dna_rm.chromosome.20.fa_hisat

Then you can perform alignment and generate a SAM file, here analyzing data from the well-studied cell line GM12878:

$ hisat -x ../reference/Homo_sapiens.GRCh38.dna_rm.chromosome.20.fa_hisat --sra-acc SRR038448 -S myAlignment1.hisat.SRR038448.GRCh38.chr20.sam

The -x argument points to your reference genome, the --sra-acc argument is to specify the SRR run accession, and -S is the name of your SAM file. Typical next steps might be to use samtools to compress your file to BAM, sort, and index; then use stringtie to assemble gene models and quantify expression; and perhaps IGV to visualize your results.

This example is from an RNA-seq lecture in the ongoing NCBI NOW workshop (Biostar tags linked here). Good luck!

That ends up downloading the whole thing.

Yes you're right, you do need to download a reference genome and you do generate a SAM file locally. But the advantage of using HISAT with NCBI-NGS toolkit is that you don't need to download reads in the FASTQ format and those are often large. So it's not quite downloading the whole thing and I believe it does address the question that was asked.

You not only download the reference genome but you download at least most of the SRA file (it doesn't write it to disk, but it still has to download it). That is smaller than the fastq files, at least :)

Log in to answer this question.