Differentially expressed isoforms with Miso

[ninni]

Hi all,

I was recommended Miso for alternative splicing analysis. Now, I have a question on how to do this in comparison with two groups (of ten samples each).
I can only calculate the Psy value for one sample at a time, but how would I pool or compare the values in order to find isoforms that are expressed in my cases?
compare_miso only works with two samples, but I would like to compare all ten versus all ten of the control.

Thank you!

[gtduarte]

Hello ninni,

I am also figuring out how to run MISO. Many ppl recommend it, but even the installation is not trivial and also not easy to find solutions for the errors.

In my case, I have two groups (treated and untreated), with 3 paired-end replicates each. However I am still testing the best options for mapping with TopHat and computing psi values, so I am using just one paired-end sample from each condition.

What I am intending to do is to run the mapping for each sample and sort them with Samtools. Then, extract the common part of the headers from these files (samtools view -H) to a txt file and use it to create a merged BAM file (samtools merge -h headers.txt merged_file.bam sample1.bam sample2.bam sampleN.bam). These merged files must be sorted again (MISO will ask for the BAI file). Since I have PE reads, I must also estimate the mean insert length and SD of the merged files before running MISO. Then, compute the psi values for each merged file and run the comparison. Hope it works

Please someone correct me if my idea of merging is not the best!