Christine,

Its very relevant question. As somebody who worked on meta-analysis of RNA-seq data, I have faced similar questions.

Now as per your question, you are looking to identify differentially expressed exons and eventually transcripts. So I feel that you need to download the sra files. The reason being the bed file that most studies provide could be raw/normalized read counts for specific gene/transcripts. Once the read counts of all exons are merged into transcript/ gene associated read counts it is impossible to obtain expression level for each exon.

If I was in your position, I would download sra files related to study of interest, then use SRA tool kit to convert sra file into fastq and then use tophat to align. The aligned and sorted bam file can be loaded into IGV to visualize the exon usage. IGV includes sashimi plots which give clear idea of exon usage in specific condition.

This method will also allow you to draw consensus from multiple studies and come to reliable conclusion.

Thank you Darshan, this gives me a clear path forward that I will dig into. I agree that the exon count is really what I am after. So, to summarise:

-download the sra files

-use sra tool kit to convert them into fastq

-align them with tophat (?bam files)

-load bam files into IGV

-look into sashimi plots

Each step will have a steep learning curve but at least, now, I have a flowchart to follow. Thank you for your reply, I am very grateful.

Kind regards

Christine

Hi Both, I had reanalyzed this scRNA-seq data, I remember that I had to analyze this data from scratch,

I would recommend to map with STAR, I invested significant time to figure out that STAR worked better on these single cell datasets

Thank you, I'll use star instead of tophat.

Thank you again for your comment.

Christine

Hi All

I thought I would follow up on my own question. Thanks to the suggestion to use the SRA file, I set out on my mission to align the RNA seq to specific exon of my gene of interest. Here are the steps I took (RNA seq for dummies)

-in geodata, I realise I can click on the GSM # of a particular sample

-this leads me to a page which under 'relations' gives you the famous SRA # (SRX)

-clicking on this, we are transfered on the run # (SRR)

-clicking on this, we arrive at the sequence read archive

-under metadata, there is a subsection with the experiment and a blast button which takes you onto a normal blast page where one can paste the sequence of the specific exon of the gene of interest in 'querry sequence'

-one can fiddle with the 'algortith parameters' section and select a high 'max target sequence' to display all the rna seq aligned

-click on 'blast'

The utube video that is quite useful is https://www.youtube.com/watch?v=FjYO6Ys5cpc

Thank you everyone to get me on tracks!

I am very grateful

Christine