multiple samples differential expression analysis

Question

[julie_dddd]

I have all these different samples and I want to do differential expression analysis: see which genes are differentially expressed in samples from healthy vs. ulcerative colitis patients. I use following samples: https://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-7860/samples/

However, I don't know if I need to generate fastq files, each time, for each sample, which would be: fastq-dump ERR3290955 fastq-dump ERR3280947 etc.. for all 144 samples

Or do I generate 1 fastq sample for a run? which is then fastq-dump ERR3280962

Since I'm very new to bioinformatics, I don't see (for now) how it differs and what we basically need to use to do differential expression analysis...

Thank you very much in advance.

Answer 1

[swbarnes2]

The typical way to work with data is to have one fastq per sample, (or a pair, if the sequencing is paired) which makes one bam per sample, make one gene count file per sample, then combined all the gene counts together, like with tximport.

[julie_dddd]

Okay, thanks! I understand. But how do I know which accession number I take, because there are ERX (for experiments), ERS (for samples), and ERR (for runs). Now, I took 3 ERX (ulcerative colitis) and 3 ERX numbers (normal patients). Are there major differences or how do I need to interpret this?