Calculating Reads And DE From BAM Files

Question

[pjmaguire3]

So I am trying to process some RNA-Seq data. I have built an annotated reference genome and mapped my reads against it using STAR, outputting the results into sorted BAM files. The issue I am facing right now is quantifying the read numbers, with subsequent calculation of differential expression(DE). After reviewing a lot of literature and posts on here, I decided to go with RSEM for calculating expression values and Limma for determining DE.

The issue I am having is that I can't figure out how to get RSEM to quantify my expression values from the sorted BAM files. Full disclaimer, as it stands I am just giving it the annotated reference genome I built using STAR. Does it really want me to rebuild a reference genome just for quantifying counts? Because while that is not the end of the world, that does not seem like a good use of resources for such a relatively simplistic task.

Example code:

rsem-calculate-expression -p 8 --time --bam --paired-end /home/peter/Documents/Files/Output/s_A/Aligned.sortedByCoord.out.bam /home/peter/Documents/Files/STAR_Reference s_A

Thank you for the all the help! If you have a different recommendation on a pipeline feel free to share it. I have two(2) replicates per sample, in case that influences anything. I am so close to finishing this project, so it is frustrating to be held up near the finish line.

[Sam]

You can always play with featureCounts or HTSeq which can also gives you the count from your bam file. For RSEM, AFAIR, RSEM requires the alignment to be done on transcript level and most likely will require you to perform the alignment again, though RSEM does support STAR, so you can use STAR for RSEM alignment with the --star option.

To conclude, if you don't want to perform the alignment again, you can try featureCounts or HTSeq

[pjmaguire3]

Thank you for the link to why it wants to run it alignment itself. In the future, if I want to use RSEM, I will make sure to run STAR through its pipeline. As for my current project, featureCounts ended up working out perfectly for my needs. Thank you for that recommendation!

[1769mkc]

I was looking for the same since i m looking for transcript level quantification so I aligned the file in quantmode in star now i have bam files so i can simply use featurecounts and do differentiatial exon analysis ?