Tweet
Dear SEQanswers community:
I hope this is not a duplicate of earlier threads, but I was wondering the following:
We have been using the Tophat2 -> htSeq -> edgeR/DESeq2 pipeline for differential expression analysis of human RNA-seq data for several years. We typically align our fastq files to a human genome build, then use a suitable GTF file to assign feature counts prior to differential expression analysis.
A colleague recently suggested that it might be better to map to an annotated human transcriptome, rather than a genome build, in this pipeline.
I was led to believe that mapping to the genome was generally preferred for differential expression analysis. I also thought that assignment of feature counts using an annotated GTF file, in effect, focused the reads around the transcriptome.
I have seen many papers examining the relative merits/problems with mapping to the variously annotated transcriptome resources (e.g., ACEVIEW, Ensembl, RefSeq, etc.), but I have not yet come across any papers in which these were compared to genome matching and feature counting using htSeq.
I was wondering if I might be missing some key resources? As well, I was wondering if members of the community had any opinions on this issue, based on their experiences. Has anyone tried such a comparison?
Thanks in advance for any advice you might be able to provide.
I hope this is not a duplicate of earlier threads, but I was wondering the following:
We have been using the Tophat2 -> htSeq -> edgeR/DESeq2 pipeline for differential expression analysis of human RNA-seq data for several years. We typically align our fastq files to a human genome build, then use a suitable GTF file to assign feature counts prior to differential expression analysis.
A colleague recently suggested that it might be better to map to an annotated human transcriptome, rather than a genome build, in this pipeline.
I was led to believe that mapping to the genome was generally preferred for differential expression analysis. I also thought that assignment of feature counts using an annotated GTF file, in effect, focused the reads around the transcriptome.
I have seen many papers examining the relative merits/problems with mapping to the variously annotated transcriptome resources (e.g., ACEVIEW, Ensembl, RefSeq, etc.), but I have not yet come across any papers in which these were compared to genome matching and feature counting using htSeq.
I was wondering if I might be missing some key resources? As well, I was wondering if members of the community had any opinions on this issue, based on their experiences. Has anyone tried such a comparison?
Thanks in advance for any advice you might be able to provide.
Comment