Open sschmutz opened 3 years ago
Commit 0edb6c2 on branch issue-6 applies a quick fix of this issue but it would be better to find the cause of it and fix that.
This issue would by the way also fail to create a wobble-free sequence if one wants to create a "majority consensus" sequence by just setting the variant_threshold to 50% (which is in my opinion not a very robust approach anyway as a major base could always also be present at <50% if there is more than one alternative base for example).
I found that this issue is caused by CRLF line endings in wrapped reference fasta files (instead of LF). The reference sequence of UL97 contained these, while the reference sequences of the other two genes we are looking at do not.
Using either unwrapped fasta files or ensuring LF line endings avoids this issue.
PS: The reference sequence should also not contain any ambiguous nucleotide codes as this also leads to issues for the consensus sequence.
Strangely affects CMV amplicon of gene UL97.