STAR BAM into RSEM

Question

[Biogeek]

Hi,

Usually when I used RSEM I feed my left reads and right reads in as .fastq files. Left is the forward and right is the reverse. I took genome files created an index in STAR and uses quantmode so that I can put the .BAM out file into RSEM.

In the STAR paper it says "--paired-end --forward-prob 0 options are applicable to paired stranded RNA-seq data such as Illumina stranded Tru-seq protocol. Refer to RSEM documentation for detailed description of RSEM parameters."

When I refer to the RSEM manual, it gives this -

----------------
--forward-prob <double>
Probability of generating a read from the forward strand of a transcript. Set to 1 for a strand-specific protocol where all (upstream) reads are derived from the forward strand, 0 for a strand-specific protocol where all (upstream) read are derived from the reverse strand, or 0.5 for a non-strand-specific protocol. (Default: 0.5)
----------------------

We used Illumina sequencing, however forward reads are always left and reverse reads are always left.

This has left me confused. Can someone explain this to me kindly?

Thanks.

Answer 1

[Devon Ryan]

The explanation given for the --forward-prob option in RSEM is incredibly opaque and should just be ignored. The description given by STAR is sufficient. If you have strand-specific data then use --forward-prob 0. If you're unsure, then use the infer_experiment function from RSeQC and post the output.

[Biogeek]

Thanks.

So for Illumina hiseq with truseq chemistry, we know the sense strand is the 5-3' (left: forward) and the antisense is 3-5' ( right/ reverse) read from knowledge. This is paired end with strand specific. So I'm asumming I still use an integer of 0 for this flag?

[Devon Ryan]

A stranded dataset gets a value of 0, so yes.

[Biogeek]

Thanks Devon. Just wanted to make sure. I agree with you though, the manual lacks Clarity.