I want to sequence a panel of diseases with adaptive sampling with ONT, some of them are aneuploidies. For the adaptive sampling regions for aneuploidies, we are going to include the critical regions of the syndromes but for some of them there is not much information. In order to adjust to the recommended 10% of the genome in adaptive sampling, I though we could discard telomeres and centromeres in the calculation of read depth for then assessing copy number. My first question is if this is a good approach.
The second thing is that I already have the bed file for the GRCh38 reference genome but I want to have it as well for T2T. In the UCSC table browser I'm able to easily find the centromeres positions but for the T2T I downloaded the whole cenSat and I don't know which of those regions I should leave out of the bed file.
If you can afford to sequence the entire genome (on PromethION flowcell) then do so and select for things you need afterwards. Based on some discussion we have had in slack "adaptive sampling" does not quite work the way we would like to imagine it does. It may certainly not work for the whole human chromosome.
You mean completely discarding adaptive sampling? We currently own the MinION and adaptive sampling works pretty horribly, it is not fully supported in this machine so I supposed we would have better performance with the PromethION, so we ordered that. We aim to use as many samples of different patients as possible so that's why we wanted to use AS. We also need sufficient depth to have a reliable representation of the regions of interest to be able to affirm there is a pathogenic variation.
Possibly. You already seem to have some experience with it not working. It may work better on PromethION (assuming you have access to one) but YMMV.
We just acquired the PromethION but we haven't tested it, we will do a couple of tests, thank you!