Hi there,
I'm sorry for the newbie question but I'm confusing and getting lost in my thoughts.
Having WES data, until the moment I have been using DNASeq aligner such as BWA MEM. But I'm wondering whether it makes sense to use RNASeq mapping algorithm (STAR...) or not and why.
For other hand lets say that I want to consider the alignment of the reads against only the exons of some genes. How this would affect the aligner algorithm decision?
Thank you in advance,
•
written
by
user230613
Stick with a DNA based aligner like BWA (recommended in the GATK best practises). DNA sequencing (such as WES), is sequencing DNA fragments pulled down by a pre-designed capture kit, whereas RNA sequencing is sequencing the actual mRNA. When aligning RNA reads, the aligner (STAR, HISAT2, etc) will expect reads to split across exons, as such the appropriate penalty for that is tuned - that sort of event shouldn't happen with DNA.
Thank you for the answer, it is very clear now. Could you help me with my second question, about aligning against previously extracted exons?
If you only want to consider some genes in your alignment (I'm not sure why you'd do this), then when building your alignment index, you'd need to restrict the fasta file (using BWA-MEM as an example) to only those regions. So if you build your BWA index using the hg19 fasta file, you'd need to subset that fasta file to only the bases in genes x, y, and z, instead.
Overall though, you shouldn't really "cherry pick" loci to align against, you should be completely agnostic, and let the reads align where they align.
The idea of using an RNAseq aligner is to be able to align across intron-exon junctions. Your exome data consists of captured exon sequences (+ maybe a little intronic sequence, depending on the probes), which are not supposed to be spliced together. Since you don't have to align across junctions, an RNA aligner will not have an advantage over a DNA aligner, but I don't see why you couldn't use either. If you only want to align against exons (i.e. each exon is an individual sequence), this still holds, as -- again -- you do not need to map across any junctions.
Log in to answer this question.
