Remove 6183, 9142 and 13339 from masked sites

[xz-keg]

NB.1.8.1 is becoming dominant in many places, however these masked sites make it difficult to analyse it as it will force seqs to adapt these positions of JN.1 which in fact do not appear in NB.1.8.1, causing newly-uploaded seqs to have artificially reversions on these sites.

Also analyse becomes increasingly complex for NB.1.8.1-related recombinants due to these sites.

Suggest to remove them or at least remove them from XDV branch so that usher tree can correctly show mutations of XDV* seqs.

One option: remove them from masked sites and re-run all XDV seqs, hence artificial reversions of past early JN.1 that are not growing now are ignored, while real "reversions" in XDV seqs are kept.

[xz-keg]

@AngieHinrichs There are many newly-appeared NB.1.8.1/ other lineage recombinants now.

In the recent GBW batch, there is a new LF.7.9.1/NB.1.8.1 recomb, a JN.1.11.1/NB.1.8.1 recomb, and an LP.8.1/NB.1.8.1 recomb. With this NB.1.8.1 usher problem it becomes very hard to distinguish and analyse them.

This becomes an urgent problem that needs to be fixed to analyse NB.1.8.1-based recombinants.