Open sebschmi opened 3 years ago
Thanks for reporting this issue, I'll investigate asap. I wouldn't be surprise that that the issue is related to the very small k-mer size, I don't remember testing that edge case. Did you have the same issue with larger k-mers (k=5,6)?
Thank you for the quick answer.
Thank you for the quick response.
I get the same issue for k=30 with the following input reference: CCCCCCCCCCCCCCGCGGGGGGGGGGGGGGG
(13 Cs attached to the front and 13 Gs attached to the back of the original string)
It is not possible to create the same test-case with uneven k, as then there would not be any self-complemental nodes.
Consider the following input reference file:
a.fa
Using the command line
Bifrost build -k 4 -r a.fa -o a
on the current master of this repo yields the following result:a.gfa
The reverse complement of
S1
isCCGCG
, which does not overlap in three characters withCGCGG
. So the edgeL1
must be wrong.Note that the node
CGCG
is self-complemental.