John Lees
John Lees
Talking to Tommi, he suggests that better compression and working on the compressed structure may be preferable. Some ideas: - The split k-mers are difficult to compress. A tree structure...
Tommi suggests: - Static bitvector of length 4 to encode each base in variant matrix, whole matrix is then a Vec of these (with correct slicing to get the index)...
See branch `const-refactor` for a prototype But for this to work properly it needs to either be bit packed and not many vectors, or a sparse array
I'm guessing that if you assemble the data first and use the fastas with `ska build` you might get decent results, but using the reads directly almostly certainly won't work...
Closing as stale
Thanks for the testing and files. The `cov.txt` file is actually fine, the coverage/error column is actually just to assist with the plotting. It looks like the default mixture model...
Not going to do this unless a few people ask for it
@sydelstan opening new issue to track this I'd require more information to diagnose this: command and ideally a minimal reproducible example (e.g. a genome that causes this issue)
I have checked with my colleague @HarryHung in the GPS project who says the following: > Checking ERR884352 > It is one of the GPS samples with: > Public Name:...
It's a good idea, I did this in sketchlib.rust with the `--concat-fasta` option. Code here if you want a reference to give it a go https://github.com/bacpop/sketchlib.rust/blob/main/src/sketch.rs#L283-L329