How to Align only few transcript and get count?

Question

[John]

Hi, I have Two questions.

I did RSEM for my (single-end) RNA seq reads. For a gene, I got isoform count for only one transcript variant. But I see three transcript variants in NCBI Nucleotide search. May I Know Why? (I used mm9 reference transcriptome for alignment using bowtie2 within RSEM)

So Now, If the reference I used was wrong, I don't want to do alignment for whole transcriptome reference again. How can I get counts by aligning only three transcript to my reads?

Thanks in anticipation!

Answer 1

[swbarnes2]

Just because NCBI shows that three transcripts exist doesn't mean all three must be present in your sample. Are they in your reference transcriptome? Does IGV show that the other transcripts are present?

I don't want to do alignment for whole transcriptome reference again.

Well, you should. It's the safest thing. If only because it's much easier to document that you did things the right way than to document how you deviated from that.

[John]

It's not even listed in the isoform list (there's only one isoform is in the result table).

Thank you

[Kristoffer Vitting-Seerup]

Which transcriptome database did you quantify with RSEM in the first place?

[John]

UCSC ...................

Answer 2

[John]

What I found: We should use Refseq reference genome and Gff3 annotation file for alignment. SO we will get Refseq Transcript ID details in here.