I've sequenced targeted DNA panel. I'm aligning it using bwa mem once to the whole genome and twice to the targeted sequences. When aligning it to the reference genome I'm getting ~30% of the reads mapped to the targeted regions. However, when aligning it to the panel sequences 70% of the read are aligned successfully.
Any idea how can it happen? even if there is favored alignment in the whole genome these reads should be assigned as supplementary alignment? Isn't it?
Hello,
there are only secondary alignments (not supplementary) if the probability of these alignments are similar. So in your whole genome alignment lots of your reads maps significantly better to other regions, than the one you guess.
That's a very good example why you should always align to the whole genome, even if you use panels.
fin swimmer
Yes - I agree.
In fact, further filtering the variants and/or alignment to just the target regions (after initially allowing reads to be aligned to other regions) should help with reducing false positives.
Even if the capturing oligos were unique/specific in silico (which they typically are otherwise the kit or design would be of poor quality), this does not ensure perfect capturing efficiency and absence of off-target captures during the experiment. Therefore, as you say, align against the entire reference genome.
Log in to answer this question.
Thanks for all responses.
I understand why aligning to the whole genome is recommended.
However, I accepted that most of the reads will still align to the targets when aligning to the whole genome but will be assigned as secondary. It seems that this is not the case here....