Marcel Martin

See #685 - [x] Log how long it took to create an index - [ ] Log how many adapter sequences were *not* included in the index (and possibly why)...

_Originally posted by @eboyden in https://github.com/marcelm/cutadapt/issues/685#issuecomment-1484470430_: > I knew we were building an index but didn't initially see the message because I was running `--report minimal`. I turned that off...

Use poolstl to sort randstrobes in parallel (attempt two)

13

We tried this before in #375, but one reason for closing the PR was that it was not possible to use a custom sort function (`pdqsort_branchless` in our case). @alugowski...

I just added a `pluggable_sort()` that makes this easy: ```c++ poolstl::pluggable_sort(poolstl::par.on(pool), randstrobes.begin(), randstrobes.end(), pdqsort_branchless); ``` _Originally posted by @alugowski in https://github.com/ksahlin/strobealign/issues/375#issuecomment-1890823659_

Use ksw2 to align soft-clipped ends from ungapped alignment

7

This should fix some unexpected alignments as reported in #357. When extending a seed, we currently compute an ungapped alignment allowing softclipping and report this unchanged if the hamming distance...

Try extension alignment if seed appears to be ungapped

1

Btw, what does strobealign currently do when the region with the NAM has the same length on the query and the reference and Hamming distance of the NAM region is...

Structure of aln.cpp

1

This shows which functions call which other functions in `aln.cpp`. This is an updated version of #98. I’m documenting this because I think some functions should be renamed to make...

As discussed previously, we should try to switch to a 2-bit encoded reference to save memory. Any nucleotides that are not A, C, G, T would need to be mapped...

Duplicate alignments lead to spurious multimappers

2

This happens when mapping single-end reads (I have not tested paired ends). Sometimes, the list of NAMs contains different NAMs that lead to the same alignment. This is visible when...

Optimized parameters lead to reduced mapping rate

1

This was being discussed in a closed PR, so here’s a proper issue for it as I am not sure what the conclusion is. PR #345 introduced syncmer/strobemer generation parameters...