The extra benefit of refining common key blanks against locks that don't change can add some resilience, but neither process can predict or pre-empt future viral mutations. /66

So although Matrix-M's strength in encouraging more apprentices to examine more locks (epitope spreading) is real and useful, it still only works with the locks actually present in the vaccine. Completely new or significantly different locks will still require new learning. /65

... while public antibodies relate more to improving the quality of keys for locks already being worked on, with the occasional bonus that an improvement to a key might remain useful against future changes to a lock. /63

This is different from epitope spreading, which is about getting more apprentices to notice and learn about additional secondary locks that are already on the Spike, including the less obvious ones ... /62

This refinement is not completely random. Improvements that make the key work better on the locks actually present will be kept. But occasionally, one of the useful refinements also happens to let the key still turn other locks that have been slightly altered on later safes. /61

... type of lock (often ones that tend to stay similar across different safes), it might give those apprentices a head start. While the apprentices learn about the locks, the coaches help them carefully file and refine the blank key so it fits the lock tightly and precisely. /60

These keys are known as “public” antibodies based on standard templates that many apprentices can produce. Because the key blank already has a rough shape that corresponds to a particular ... /59

But there is another process not specific to Nuvaxovid that can provide some extra resilience. Some apprentices might start with a very common, ready-made key blank that a large share of apprentices already know how to work with. /58

We discuss a real-world example of this more in the “Effectiveness” section below. If a later version of the Spike protein were to arrive with changed locks (mutations), the B cell safe-crackers will need fresh coaching to study the new locks and learn to make keys that fit. /57

However this has limits. A vaccine can only train the immune system on locks (epitopes) actually present on the safe (Spike protein) seen. It cannot predict the future or invent keys for locks that have never appeared. /56

After 2 doses, neutralization levels (titers) for the original COVID variant were about 48 times higher than against BA.1 Omicron and 34 times higher than against BA.5.  By the 4th dose, this gap narrowed  significantly to 6.5x and 3.6x, respectively. /54

On the map, this is shown by a shorter distance between the original strain and Omicron variants. In other words, the antibodies became better at bridging the gap between the differences between variants. /53

Neutralization against the later Omicron BA.1/BA.5 variants was much weaker than the original Wuhan strain (the points sat further away on the map). After a third dose, and more so after a fourth , the same people's antibodies neutralized Omicron variants relatively better. /52

Each square on the grid relates to roughly a two-fold neutralization drop. Variants that sit close together are recognized similarly while those further apart are recognized less well. One study looked people with 2 doses of the original Nuvaxovid (2020 Wuhan strain). /51

This is seen with antigenic cartography, a way of mapping neutralization measurements visually. Researchers measure how well a person's antibodies neutralize different variants and plot the difference. The distance between points reflects the difference in neutralization. /50

... plenty of the coaching needed to expand and mature and generate keys (antibodies) to the secondary locks (epitopes) on the safe that don't have any keys made for them yet. /48