What human transcriptome annotation to use for RNA-seq analysis

Question

[guillaume.rbt]

Hello,

Sorry for this naive question, but I'm starting a RNAseq differential expression analysis, and it's my first time with human data.

I'm wondering which annotation should I use for my analysis, is there a "standard" annotation that is efficient and commonly used? Should I use an annotation of hg19 or hg38 assembly?

Thanks

Answer 1

[caggtaagtat]

Hi,

I would use the improved and newer one hg38

Answer 2

[Benn]

I use the ensembl annotation, which you can download here (Gene sets: GTF).

[guillaume.rbt]

Thank for your help. As I will be using Salmon for quantification I need the sequences of all transcripts, should I get the "cDNA" Fasta for that?

[GenoMax]

You could do that.

Alternatively, GTF files and fasta files for transcripts are also available from GENCODE site (where Ensembl data comes from).

[guillaume.rbt]

Ok thanks for the information. Would it be wiser to use the "Transcript sequences", or the "Protein-coding transcript sequences" from GENCODE? (as I will be analysing expressed gene I think it's ok to use only the Protein-coding transcript sequences, but maybe I'm wrong)