SNAPP & SNAPPERS management of missing data

[Loukia Spilani]

Hi BEAST users! I have been trying to build and date my species tree by following this tutorial: https://github.com/mmatschiner/tutorials/tree/master/divergence_time_estimation_with_snp_data#q2.

I have a ddRAD dataset, so naturally, some individuals have worked better than others and there's a considerable variation is missingness percentages. Additionaly, not all the species include the same amount of specimens/lineages. Because of these two things, there are SNPs (I would argue most of them) that are only present in some species. Given that snapper is based on snapp, and the information on this post: https://www.beast2.org/2014/07/21/snapp-handling-missing-data-and-path-sampling-made-easier.html ("... no sites with missing data are deleted any more, even when this leaves some species having zero lineages. SNAPP assumes the gene tree simply has no taxa in that species any more.") , I was assuming that both SNAPP and SNAPPER can handle species for which all individuals are missing a specific SNP. However,  when I started running snapper, all my SNPs got filtered out (WARNING: removed 1000 patterns (1000 sites) because they have one or more branches without data.)

I know that the post is relatively old (it refers to SNAPP v.1.1.6 while I now have version 1.6.1), so I was wondering if the handling of missingness is still the same as what is described in that post, or if it has become more stringent with time. Or maybe I get this warning not because of missingness per se, but maybe because the selected 1000 SNPs are not informative, so they are all discrded.

Thanks in advace for any info, and sorry if this has already been addressed. I tried going through the group and github and the only recent mention that I found regarding snapper was in this thread https://github.com/rbouckaert/snapper/issues/3 but it doesn't clarify how missingness is handled, it just says that it acceptable