Program For Flagging Alternative Splicing Events From A Genome Annotation

Question

[Daniel Standage]

I'm looking for a program that will take a GTF/GFF3 file as input and produce a report describing the observed alternative splicing events (skipped exons, retained introns, mutually exclusive exons, etc). Honestly, it seems like a very straightforward task to write such a program/script, and I was planning on doing so myself. Before writing the code, however, I thought I would make sure that nobody else has shared code for doing this type of task before.

[Rick Meyer]

I was also interested in such a program. Have you found anything or would you care to share what you developed?

[james.lloyd]

I am now looking for the same thing. To take a Cufflinks GTF and annotate each transcript with splicing type. Why is this so hard?

[mfahim]

did you find anything for plants?..

[james.lloyd]

Yes I found SUPPA. It worked well. It is a set of python scripts and the generateEvents script is what you want to run on the GTF from Cufflinks/Cuffmerge and then the output files called .ioe have the events.

Answer 1

[pengchy]

Hi Daniel,

How about your work on this topic going on?

I have found there are two package worked for this purpose, i.e alternative splicing event classification.

1. LESSeq [1, 2], this package uses the MRF format to store the alignment developed by Gerstein lab. Not yet published and the documentation is scare.
2. spliceR [3]
3. spladder [4,5]: The input annotation file should be in cpickle format.
4. rnaseqlib [6]: the input is genePred file

---

• [1] https://github.com/gersteinlab/LESSeq
• [2] Gerstein, M. B., et al. (2014). "Comparative analysis of the transcriptome across distant species." Nature 512(7515): 445-448.
• [3] Vitting-Seerup, K., et al. (2014). "spliceR: an R package for classification of alternative splicing and prediction of coding potential from RNA-seq data." BMC Bioinformatics 15(1): 81.
• [4] https://github.com/ratschlab/spladder
• [5] Kahles, A., et al. (2015). "SplAdder: Identification, quantification and testing of alternative splicing events from RNA-Seq data." bioRxiv.
• [6] http://rnaseqlib.readthedocs.org/en/clip/#creating-custom-gff-annotations-for-miso

Answer 2

[Vivek]

Running Cuffcompare

Cufflinks includes a program that you can use to help analyze the transfrags you assemble. The program cuffcompare helps you: Compare your assembled transcripts to a reference annotation Track Cufflinks transcripts across multiple experiments (e.g. across a time course) From the command line, run cuffcompare as follows: cuffcompare [options]* <cuff1.gtf> [cuff2.gtf] ... [cuffN.gtf]

Cuffcompare should do this.

http://cufflinks.cbcb.umd.edu/manual.html#tracking_format

[Daniel Standage]

Two comments: first, cuffcompare is for comparison of two or more GTF files (not shocking given its name), whereas I am looking for something that will take a single annotation file; second, it's not clear to me where in the cuffcompare output I will find a summary of alternative splicing events. Perhaps you can clarify.

[Vivek]

You would have to compare the GTF to a reference to find alternative splicing events right? You get a .tracking file when you run cuffcompare that has class codes which indicate how the transcript either matches or differs from the reference.

[Daniel Standage]

I'm working with an organism that has no reference annotation. I can use TopHat/Cufflinks to align the RNA-seq reads and assemble the alignments into transcripts. In cases of alternative splicing, a single Cufflinks-produced GTF file will contain multiple overlapping transcript annotations sharing some but not all exons. Alternatively, I can use some gene annotation workflow to produce transcript annotations. Depending on the evidence I provide, this may also identify distinct transcripts that share some but not all exons.

In either case, I don't need to compare a GTF/GFF3 file to another file to identify alternative splicing events, I just need to look for transcripts that share some (but not all) exons. Am I missing something here?

[Vivek]

You could be right, I'm just not aware of how to go about finding alternative splicing events for non-model organisms without a consensus reference to compare to.

If I had to take a crack I'd probably consider the transcript with highest abundance as the reference transcript and annotate alternative splicing events by comparing the rest to to it. That's just my guess.