PDFs and HTML output in miRDeep2

[Former user]

Hello, I'm trying to use the miarma pipeline for miRNA novel prediction but it seems to happen that when miRDeep2 module runs, the PDFs and HTML output results from make_hmtl.pl after quantification of predicted known and novel miRNAs are somehow missing.

While running miRDeep2 as independent program, not inside miarma pipeline, those formats are quite valuable (mostly PDFs with predicted structure and aligned sequences in samples).

I can assume that somehow the -d option in make_hmtl.pl script inside miarma (which disables PDF creation) is activated. Could it be any option to modify the miRDeep2 module inside miarma to allow PDFs to be created?

Thanks

[Eduardo Andres Leon]

Dear user, thank you for using miARma-Seq

About your question, the PDF and HTMLs outputs are disable by default within miARma. When developing miARma, It was tested in seven different systems and the HTML::PDF perl library needed by miRDeep2, was incompatible among them. So we decide to disable the PDF creation.

If you think that these files must be included, I can try again to find a solution to these perl libraries. Let me know what you think, in order to change the type of your issue from bug to proposal

Thanks in advance.

Eduardo.

[emarmols]

Oh sorry you are right, this issue may be better placed in proposals.

I've tried miRDeep2 with my set of samples and wanted to compare results if consistent. For me it would be of great help to have the -d and -v options disabled in order to have a proper overview of all that is going on there within the workflow.

I tried modyfing perl libraries but could not find where those variables are defined by default.

It could be nice as well if adapter trimming in De_novo Module could be optional, say, being able to load already trimmed samples and avoiding trimmer from within miRDeep2 to do it. Getting already installed cutadapt program to do it as an option inside De_novo Module instead of launching automatically te miRDeep2 trimmer from raw reads.

Thanks in advance

[Eduardo Andres Leon]

It's not a problem, I've changed already.

I guess you are looking for the line 2179 in the Aligner.pm module. But the line 2217 must be also modified (commented for example). The first one is where mirdeep is executed. The second one removes all temporary files from miRDeep.

I managed to share the HTML::PDF libraries among debian and ubuntu, but no success yet with apple/Fedora and CentOs.

I'll let you know

[emarmols]

Thank you very much, those were the lines I was looking for to modify. I'm using an Ubuntu 16.04. I will let you know.

[emarmols]

Dear Eduardo, I tried modifying the perl script called Aligner.pm module in 2179 line and it all worked ok, PDFs were generated properly. I found that modifying line 2217 by commenting it resulted in an error of execution and final count archives by featureCounts were not generated though, so I kept this line as it was. Just removed -d option from line 2179.

I also kept wondering why did you put -c option (which disables RANfold analysis). I assume it was put there in order to save computational time, as RANfold analysis takes a good amount of time for each sample to be completed. Just wanted to know if there were any other reasons behind this ultimate decision or it was just a time saving trick.

Thanks

[Eduardo Andres Leon]

Dear Emarmols, You are complete right. RNA fold is disable by default, but we provide RNAfold binary because in future release we will let the user activate this option. Once we solved the PDF::API2 problem, -c and -d will be offered so, you will be able to activate it or not from the ini file.

We hope to provide this version asap

Thanks for your very appreciated suggestions

[Eduardo Andres Leon]

Issue #67 was marked as a duplicate of this issue.