Covid patients can remain infectious following recovery.

[NNadir]

13. The mechanism by which resistance arises in bacteria, and in viruses is similar.

It is actually most widely observed in viruses, since, while bacteria replicate rapidly, viruses to do even faster.

It's all about the Nietzschean "That which doesn't kill you makes you stronger."

Early in the AIDS epidemic a large class of AIDS drugs were developed in a class of drugs known as Protease inhibitors. If one took one of these, and most patients started with just one of the available drugs, they were effective at deactivating the virus above a level we call a "trough." The dosing schedule is designed so that patients will have physiological levels well above this trough. But if one forgot to take a dose, one could fall below this trough.

Most mutations result in inoperative viruses that cannot function. But in rare cases there are mutations that allow for a functional virus that has a mutation allowing for continual functionality and improved resistance to a drug.

Consider that viruses replicate on a scale of billions of viral particles a day. Suppose, because the machinery of the virus replication is primitive, that 100 million are mutants, most of which are no longer functional. But in that hundred million there are say, 500 that are functional, and of these, there are 10 that are functional and have developed some level of resistance to the drug. As the amount of the drug goes lower, many of the viruses are rendered inoperative, but those 10 have enough resistance to overcome the effects of the drug. There is nothing to stop these survivors from infecting cells, since they are functional. If they are allowed to do so and are not killed, all of their progeny, except those damaged by poor replication, will come to dominate the viral load by natural selection.

We have fully sequenced the target HIV protease, but as we observed patients showing resistance, we found sequences that were different than those for which the drug was designed.

Protein sequences are often described by one letter codes representing each of the 20 eucaryotic coded amino acids.

Here is a list, from my files of mutants to HIV drugs:

D30N: Nelfinavir. (Agouron/Pfizer).
M46I/I47V/I50V: Amprenavir (BMS).
L10R/M46I/L63P/V82T/I84V: Indinavir (Merck)
M46I/L63P/A71V/V82F/I84V: Ritinovir (Abbott).
Saquinavir: G48V/L90M (Roche)

D30N means that at residue 30 in the protease, an aspartic acid (D) has been replaced by an asparagine (N). (This old stuff by the way, dating back well over a decade.) Notice that in some cases, indinavir and retinovir, there are multiple mutations.

This fast evolution is driven by high replication rates coupled with large replication inaccuracy. Now it may be that a particular single mutation does not confer full resistance, but allows for enough resistance, to allow for the virus to "survive" when the level of the drug falls below a certain level because the patient skipped a dose, or because the drug was not available because of costs or logistics.

Since sub-saharan Africa is an AIDS hotspot, and because the drugs were sometimes only sporadically available, this made for rapid mutations. If a patient does not transmit these variants to another person, it stops with that person. But that is not always observed, as we know.

I hope this helps explain the issue.