Glitch or Feature? Either methods do not represent the true distribution of an allele when gaps are involved

[moa4020]

Hello! I use ggseqlogo to produce logoplots for HIV variants. I recently have come across an inconsistency that I have questions about.

Here is an explanation to the issue I have: In Donor#X, a Gap in site 26 is not reflected on the logoplot. ggseqlogo shows a gap in that site only if all the sequences in that position have a gap. Or else it gives a 100% probability to an allele even if it's present in just one sequence. This was seen in both

Here is an Image I have attached of the logoplot and the consensus matrix. Please see sites 26 and 27.

The A at 26 is full sized even though it is present only in 1 sequence out of 18 total sequences. The rest of them have a "-".

Similarly, I expected site 27s G to be a little smaller due to the fact that it was seen only in 17/18 sequences.

I got the same results even with the "Bits" method.

methods.

[omarwagih]

Can you share example data for me to be able to fully reproduce?

[moa4020]

OUT.zip

I have included the aligned fasta sequences and my consensus matrix and logolot for reference.

[moa4020]

Hello! Just following up to see if there are any insights to this issue.