Enhancement request: Implement +ASC option to speed up MixtureFinder searches on supermatrices lacking constant sites

[vinuesa]

Dear IQ-TREE developers, I have been testing MixtureFinder and found that it runs significantly faster on large concatenated datasets after removing the constant sites. However, the versions I have tested (v2.3.1 and v2.3.2) do not implement the ascertainment bias correction (+ASC) option that would be desirable to avoid potential branch length overestimation. Could this option be added to a new release? Thank you very much.

[bqminh]

perhaps @HuaiyanRen can accommodate this request?

I'm however not sure how many users would be interested in such feature. We'd need to prioritize tasks among many things we want to do...

[vinuesa]

Thank you Bui @bqminh and @HuaiyanRen for your consideration. Your point is clear. I'd just like to let you know that I'm implementing the MixtureFinder option in our GET_PHYLOMARKERS pipeline, and therefore the user base would increase somewhat. Best regards.

[roblanf]

@vinuesa, that sounds great. W.r.t. enhancement requests, the more support you can demonstrate from potential users (e.g. via thumbs up on the original post, additional comments, etc) the better. We have quite limited time to implement things, and the number of interested users is a strong consideration.