Bernadetadad
commented
9 months ago This would of course be much easier to confirm if I had good BA.2 serum mapping, sadly most of it due to spike in issues is not great. But for very potent sera where VSV-G neutralization was likely less of an issue like it really does seems that these sites that we would think important for RBD modulation have far less/no effects. E.g. https://dms-vep.org/SARS-CoV-2_Omicron_BA.2_spike_DMS_sera_and_mAbs/405C_escape_plot.html
jbloom
@jbloom one reviewer is asking for explanation why correlation between yeast DMS and XBB.1.5 ACE2 binding is significantly worse than that between yeast DMS and BA.2 ACE2 binding.
I think it has something to do with XBB.1.5 RBD itself, specifically, I wonder if somehow XBB.1.5 RBD is more "sensitive" to the binding changes caused by RBD movement. The reason I wonder this is because XBB.1.5 and BA.2 library measurements don't have great correlation, but XBB.1.5 full spike and XBB.1.5 RBD only library binding correlated really well. I also think that a lot of sites that are off diagonal in these correlation plots are the ones that we would think are involved in modulating RBD movement - which would explain why correlation with yeast data is not great and if my hypothesis is correct would also explain why correlation with BA.2 library is not great.
To that end, I wonder if we could make a correlation plots between XBB.1.5 full spike library ACE2 binding and BA.2/yeast ACE2 binding and colour the point by R values for ACE2 binding vs serum escape? This might need to be filtered to show only sites that has some minimum number of mutations measured otherwise it may be quite noisy.