-
Notifications
You must be signed in to change notification settings - Fork 36
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
A few questions about my application (changing attributes, ultralow endsfree, endsfree API, max_score heuristics) #93
Comments
Hey, sorry for closing without a message. I was just about to write this explanation. The scope of my project shifted a bit and for now, I don't need to use this library anymore. But I'm still interested in its development and happy to discuss the above points. |
Hi, @feldroop. Sorry for not answering sooner, and thank you for your nice words. Your experimental long-read aligner sounds really cool. I am familiar with many of those ideas, and I really love the direction you are taking. If you are still interested, let's go through your list:
|
It is ok. No problem at all :-) |
Hi!
First of all, thank you a lot for providing this intriguing and well-documented piece of software.
I am working on an experimental longread aligner and currently testing WFA2 as the alignment backend. While doing so, I encountered a few issues.
As recommended, I want to reuse the
wavefront_aligner_t
objects. However, I need to change some of the attributes before each alignment invocation. Which, if any, of the attributes are safe to change on thewavefront_aligner_t
object (i.e. after callingwavefront_aligner_new
)? The most important one for me are the *_begin/end_free values of thealignment_span
. Next would be thealignment_scope
, but here I could also store two aligners instead of changing. Finally, I was thinking about varying thememory_mode
depending on the size of the input sequences. From looking at the code ofwavefront_aligner_new
, it seems like the *_begin/end_free might be safe to change after construction, the other mentioned attributes not. Is this correct?The longreads in my data sets are up to 600K bases long and I need ends free alignment. I saw in another issue that you recommended the
ultralow
memory mode for input sequences of this size. There is no urgency whatsoever, but it would be nice for me to have theultralow
mode available for ends free alignments.What I want to do is a glocal alignment, like the below example from your README. For my application, I need two pieces of information.
wf_aligner->cigar->begin_offset
to be exactly this, but it is 0 in the example (end_offset
accordingly is 50). So my question is, what exactly are thebegin/end_offset
values and how do I (conveniently) get the value I want? Is there an easier option than looking at the cigar and doing the offset calculations?cigar_sprint_SAM_CIGAR_endsfree
. I found thecigar_maxtrim_gap_linear
function, but it only trims the gaps from the end.max_score
. It would automatically use all heuristics that preserve the optimal score below the configuredmax_score
. Just an idea from lazy me :DThe text was updated successfully, but these errors were encountered: