3
$\begingroup$

I am involved in a Drosophila transcriptomics project in which we have RNAseq data for small RNAs (for microRNAs) in addition to total RNA (for mRNAs).

  • The total RNA reads have been analysed satisfactorily using the Tuxedo pipeline using TopHat, Cufflinks, Cuffquant and Cuffdiff. (I run this locally on a fast desktop Mac.)
  • Initially we applied the Tuxedo pipeline to the short RNA reads, but abandoned this because of an incidence (1 in 10) of anomalously low or high replicates that was inconsistent with the bigwig plots from the bam stage, displayed in the UCSC browser.
  • We there switched to analysing the short RNA reads using the CAP-miRSeq pipeline on Unix, which initially — though very slow — gave consistent results. However recent attempts to process later samples by this method failed as the pipeline can no longer find some dependencies it requires.

I see no point in spending further time on either of these methods.

Question

Has anyone experience of a more modern analysis method for microRNA RNAseq data, preferably one that would run on MacOS?

I have no trouble in running the Tuxedo pipeline and get good results with total RNA. The problem there is not my computer. Unless someone who has analysed microRNA using this package understands the problem I have encountered, there is no rationale basis for trying this again. As regards Cap-miRSeq, it does not run on the Mac, only on Unix. A colleague set it up on a virtual box. We are our support. Is there anyone working on RNAseq of microRNA?

$\endgroup$
4
  • $\begingroup$ Okay, thanks why not then simply use Bioconductor microRNA, which is under active/recent development? $\endgroup$
    – M__
    Feb 13 at 19:21
  • $\begingroup$ @M__ Thanks. Didn’t know about that. I’ll have a look at it. $\endgroup$
    – David
    Feb 13 at 19:42
  • $\begingroup$ @M__ I had a look at it. It seems to be a tool for identifying microRNAs. However that is not the problem. The reference database for the Drosophila genome includes the microRNAs. $\endgroup$
    – David
    Feb 14 at 13:10
  • $\begingroup$ Okay thanks, @llrs will know how to do this, but really I think you are into R-based coding. Between Bioconductor and Seurat there must be a replica of your original program $\endgroup$
    – M__
    Feb 14 at 13:55

1 Answer 1

1
$\begingroup$

Answer from @supertech, converted from comment:

This can be frustrating I understand. In my experience it's easier to read the paper and install the modern versions of the tools compatible with your computer (paper was published in 2014). Even more convenient you could open a free account at https://galaxyproject.org/support/account-quotas/ (gives 250GB free space), and do your analysis there. Tophat aligner can be very slow. STAR is better from mapping accuracy to speed; I would recommend that you use that if possible. If you are doing analysis with large genomes your laptop computer is unlikely to have necessary RAM to run the aligners.

  • Answer from @m, converted from comment:*

Why not then simply use Bioconductor microRNA, which is under active/recent development?

Looks like STAR has been used for microRNA mapping for ENCODE:

https://www.encodeproject.org/microrna/microrna-seq/

From the ENCODE pipeline developer (@user3274289):

The ENCODE microRNA pipeline uses STAR for mapping and quantification and should be runnable on mac (you might need one on the beefier side though to have enough memory but i have run some workflows on a new macbook pro). We do not have drosophila STAR index available so you need to generate that using the workflow that is provided. See instructions on github.

$\endgroup$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.