https://nextstrain.org/ncov/gisaid/global
with Delta being the large green and blue parts at the top, and Omicron being a few red dots a little up from the bottom. If you trace back, you'll see that its nearest relative is wildtype from the spring of 2020. That's pretty weird, and there are really only three ways I can think of that it could happen:
1. human manipulation in a lab using early wildtype samples. This seems unlikely, because what would be the point and how did they let it escape?
2. chronic infection in a single immunosuppressed patient. We've seen this before (see, for example, this: https://www.nejm.org/doi/full/10.1056/NEJMc2031364 ) and southern africa has a lot of immunosuppressed people thanks to HIV. Additionally, there are some genetic markers that suggest evolution in a human.
3. spillover of wildtype in spring 2020 to some animal, evolution in that population for about 17 months, then spillback last month into humans.
I'm with the article, in that I favor 3 even though there's some evidence for 2 just because i can't quite understand how the omicron patient might have gone all that time without infecting anybody else with a proto-omicron version of the virus (which would then spread and be picked up on that chart). It's possible, I guess, that omicron got increased transmissability in a mutation in the recent past, and it could have lived in the patient with bad transmissability for a long time previous. And since delta is so amazingly transmissable, any proto-omicron could have infected a couple of people and then died out undetected. But still.