How to run eXpress RNA-seq quantification?

Question

[k.kathirvel93]

Hi EveryOne,

I am trying to benchmark RNA-seq quantification tools, While trying eXpress, i donno how to finished it up coz of reference and bam file confusion. I have aligned my paired-end RNA-seq data with Hisat2 and STAR. Can anyone suggest me a code for eXpress quantification. It ll be so helpful for my thesis. Thanks in advance.

Here is my code : Am not getting any output even error.

./express -o /home/kathirvel/  /home/kathirvel/Indexed_Reference/GRCh38  /home/kathirvel/Sorted.bam

[Devon Ryan]

There have been a lot of benchmarking papers out there, is there a reason you want to perform yet another benchmark?

[Charles Warden]

While I don't exactly suggest doing benchmarking for the sake of benchmarking, I do think it is important to have some familiarity with possible methods for analysis (so, what is described in one workflow shouldn't represent your full set of bioinformatics experience).

If you have aligned reads with HISAT2 and STAR but have gotten some sort of weird result, then there may be some value in trying other options (and I believe Devon was trying to say that certain haplotype analysis may benefit from a different type of reference, such as if you were trying to assign HLA types).

Specifically in terms of eXpress, I think part of the issue is that you should have a transcriptome (or other alternative) reference, but I think you are using a genome reference.

Otherwise, if I look at a sample of code with alignment to a de novo assembly reference, I think the code should otherwise be OK:

/path/to/express --no-bias-correct [transcriptome FA] [transcriptome BAM] -o [results subfolder]

You don't have to add --no-bias-correct, but that is another option (if you have the program working). I also don't know why you aren't getting any error messages, but that code used eXpress version 1.5.1 on a Linux VM (so, I don't know if there may be some issue with your particular build).

[k.kathirvel93]

Hi Charles Warden, I am happy about your reply, and now i got some idea. Can you suggest me where can i download that transcriptome FA file. I have .gtf annotation file.

[Charles Warden]

For human, there are some FASTA transcriptome references from GENCODE.

However, you may want to filter those. Also, if you are using the analysis to address some sort of specific question, you may need to identify the most relevant sequences through some other means (which is hard for me to give you advice about).

However, in the interests of testing if eXpress works, this should be OK. Likewise, it does give you something to visualize the read counts by transcript (not just counts), if you align with Bowtie/Bowtie2 before quantification with eXpress.

[k.kathirvel93]

Yes. For specific area of disease.

[Devon Ryan]

Benchmarking results aren't going to change for a different disease area unless you have something really weird going on, like VDJ recombination.