Hi Biostars,
I am trying to understand the boundary conditions around adaptive sampling / adaptive sequencing for Oxford Nanopore.
As an example, consider the following cases:
Case 1: I want sequencing of all chromosomes except chromosome 14.
Case 2: I want to retain any reads in my sample that do not map to the human genome.
In Case 1, the amount saved is 5% or thereabouts. The amount gained by use of adaptive sampling is unlikely to provide significant enough enrichment to matter. In Case 2: depending on the sample, it is easy to imagine the target reads could constitute 0.001, 0.01, 0.1, 1% ... or another fraction.
Most simply, my question is just, "what is the range of target sequence fractions thought optimal for adaptive sequencing? At what point, on both sides, do you start to see diminished returns?
Nuancing this a bit more, I was thinking about cases like whole exome sequencing. Riddle me this: introns are anywhere from small - not much larger than exons, to huge in length. im sure that in between there are cases where the N50 of your sample itself would actually be a predictor of whether or not a given read is likely to contain exonic sequence or not.
in such cases, if you wanted to get a whole exome, youd honestly end up covering as much as 20-25% of the genome, if you take too much intronic sequence. if you do less than that, you might end up missing exons - and at any rate, you might not be able to estimate this accurately all that well, either, unless you had a good sense for your N50 going in.
so what do you do? is adaptive sequencing as a modality for whole exome basically a dumb idea for that reason (its ok if it is, i sort of suspect that tbh - dont have a "dog in the fight" either way, just trying to understand).
