Re: [Samtools-help] samtools coverage counting deletions

[Peter <pe...@ma...>]

On Wed, Jul 28, 2010 at 11:21 PM, Tadigotla, Vasisht
<Vas...@li...> wrote:
>
> That sounds reasonable, "base coverage" implying whether specific base calls are made at that particular position and "read coverage" indicating if there are any reads spanning that position.
>
> Vasisht

Is there a consensus for how to handle the "missing bases" between
to properly mapped paired end reads when calculating coverage?
Again, there are two reasonable options - just count the bases for
the two reads themselves, or also count the inferred coverage from
the region between them.

In the case of prokaryote or mitochondrial transcriptome sequencing,
with smaller RNA fragments we found only the ends of transcript got
sequenced. Here including the middle region between paired end
reads gave more visually pleasing coverage plots. The idea here is
that (since there is no splicing) if you have two well mapped paired
reads your sample contained the whole fragment between them, so
count all of it as being covered.

In the general case of eukaryote transcriptome sequencing, it is not
as useful to count the "inferred missing bases" between paired reads,
as they could be spanning introns.

Peter