suggestion for alternative splicing event detecting tools under two biological conditions

Question

[slin023]

Hi everyone, can some people recommend a tool for detecting alternative splicing events (AS) and isoforms of two biological conditions --hopefully, by two transcriptome comparisons? I see AS-quant fitting to the experiment, but it is only for mouse and human models. I am interested to see the structural aspects in exons and alternative splicing by comparing between wild type and mutant transcriptomes. If anyone can provide suggestion for the tools, please let me know, thank you!

[ATpoint]

I would check https://bioconductor.org/packages/release/workflows/vignettes/rnaseqDTU/inst/doc/rnaseqDTU.html for starters and go forward from there and explore linked and suggested strategies.

Answer 1

[dariober]

For a couple of projects I've been involved with, I have written this snakemake pipeline for detection of differential splicing. The input requirement is fairly basic: fastq files, reference genome, genome annotation in gff format.

The key aspect I couldn't find in other packages is that it detects splicing events not present in the reference gene model. For differential analysis it relies on the strategy for differential methylation from (Chen et al. 2017)[https://f1000research.com/articles/6-2055/v2] and based on the edgeR package. Roughly the workflow is:

• Align reads with hisat2.

• Use featureCounts to identify all known and novel exon-exon junctions across all samples.

• Exclude junctions supported by a small number of reads and present in only few samples. Allegedly these are transcriptional or sequencing noise.

• For each junction and in each sample, count the number of reads supporting the splicing event and the number of reads supporting the (putative) intron retention event.

• Test each junction for differential splicing using Chen et al. method. In our case, instead of counts of methylated and unmethylated cytosines at each CpG, we have counts of spliced and unspliced reads at each junction.

I don't claim the pipeline to be ideal for any experiment, at least not without some tweaking of the defaults. It's a pipeline, not a self contained package. However, it works well for us in Plasmodium and Toxoplasma. I'm happy to provide some assistance, although I cannot guarantee anything!

[slin023]

Thank you all for the suggestion and your time. The biological conditions are involved with wild type and mutant comparison, so using reference genome will be most likely different in structure-wise. That's why we conducted de novo assembly to browse the structural difference and exon positions, so my goal is not looking for differentially expression of genes. However, I just recently ran into this github pages including all tools may be helpful for isoforms analysis. Anyone is interested in related studies can take a look at it.

List of Isoform analysis tools