Quick question about rsem-prepare-reference

[Saoirse Foley]

Hi there,

I'm trying to understand what needs to be done to use RSEM for quantifying RNA transcripts in many samples. So far, I have:

- Both Bowtie2 and RSEM installed
- A reference genome for my taxon

- A GTF file

- Several different sets of paired-end reads for my taxon, each at different time points 

I'm trying to understand whether I'll need to run "rsem-prepare-reference" for each set of paired-end reads, or whether the one "reference" prepared using the genome and GTF will be sufficient when I use "rsem-calculate-expression" to estimate abundances at each of my time points.

My initial attempt in doing this was to first use Bowtie2 to generate BAMs at each timepoint, but this didn't work out because the BAMs were built based on the reference genome instead of the transcriptome. I've seen that STAR has the --quantMode flag to avoid this problem. Will I need to add any extra flags to either command to make sure this isn't a problem?

All the best!

Saoirse.

[Colin Dewey]

Hi Saoirse,

You will need to first run "rsem-prepare-reference" with your reference genome and GTF file to create RSEM's (and Bowtie2's) reference files.  Then, for each sample, you should run "rsem-calculate-expression" on that sample's reads, using the RSEM reference you created in the first step.  Regardless of the aligner you use (Bowtie or STAR), you need to run rsem-prepare-reference first.  rsem-calculate-expression will take care of calling the aligner with the appropriate options.

Hope that helps,

Colin

[Saoirse Foley]

Yep, got it, thanks!

I built the reference like so:

rsem-prepare-reference --gtf REFERENCE.gtf --bowtie2 REFERENCE.fna refSuffix

And then used:

rsem-calculate-expression -p 8 --paired-end --bowtie2 --estimate-rspd --append-names --output-genome-bam pairedEnd_1.fastq pairedEnd_2.fastq refSuffix outSuffix

And it worked. Yay!

Saoirse.