First, let me say I'm a huge fan of your work.
And apologies in advance, but I have a feature request-- namely you've hard-capped the minimum --max-depth to 10, and I was hoping you could take the guardrails off and drop that all the way to 1.
As for why, for those of us working w/ challenged samples it's quite easy to get bacterial contamination, which presents a lot like a PCR duplicate (but will not be marked as such). A very simple mitigation strategy is to reduce the --max-depth to something quite small; in the case of our <1x genomes, a max-depth of 1 may provide a reasonable tradeoff between the amount of data considered and its correctness. And in general, I can see taking an upper-bound from a Poisson if one wanted to set this filter dynamically to something sane (and perhaps <10).
Regardless, thanks for your help and dedication.
-August
(FWIW, I'm using the static binaries that you provide; thank you for that, btw! The version numbers are:
GLIMPSE2_phase v2.0.0 / commit = 3bed6d9 / release = 2022-12-07
)
First, let me say I'm a huge fan of your work.
And apologies in advance, but I have a feature request-- namely you've hard-capped the minimum --max-depth to 10, and I was hoping you could take the guardrails off and drop that all the way to 1.
As for why, for those of us working w/ challenged samples it's quite easy to get bacterial contamination, which presents a lot like a PCR duplicate (but will not be marked as such). A very simple mitigation strategy is to reduce the --max-depth to something quite small; in the case of our <1x genomes, a max-depth of 1 may provide a reasonable tradeoff between the amount of data considered and its correctness. And in general, I can see taking an upper-bound from a Poisson if one wanted to set this filter dynamically to something sane (and perhaps <10).
Regardless, thanks for your help and dedication.
-August
(FWIW, I'm using the static binaries that you provide; thank you for that, btw! The version numbers are:
GLIMPSE2_phase v2.0.0 / commit = 3bed6d9 / release = 2022-12-07
)