There is a handful of "accepted" formulae to calculate one's 1RM based on other xRMs. In fact, there's a nice Wikipedia article that lists them: https://en.wikipedia.org/wiki/One-repetition_maximum#Calculating_1RM
But in all cases, you have to keep in mind that there doesn't exist a single formula that could perfectly predict any person's 1RM given submaximal testing. There are too many variables at play between different people, and all these formulae only consider the results of one submaximal test.
Additionally, it is understood that all of these estimations get worse and worse the higher you take your repetitions in your test. Testing your 20RM is always going to give a much more volatile estimate than if you did a 3RM or 5RM because higher reps involve more muscular stamina, which isn't an inherent factor in the formula.
Furthermore, these tests are more suited for those who have trained for a while, because rookie lifters might be able to crank out 20 reps of low weight, but they haven't prepared their nervous system to even hold the weight of their estimated 1RM.
The same formulae can be used to estimate other xRMs too. To do so, you need to isolate the weight variable in the equations instead of the 1RM variable.
Here's an example using the Lombardi formula.
Originally, the equation states:
1RM = wr^(0.10)
Isolating the w variable we instead get
w = 1RM / (r^(0.10))
For this example, let's say your 1RM is 100kg, and you want to estimate your 20RM. We would insert 1RM = 100, r=20, and isolate w, giving us
w = 100 / (20^(0.10)) ~ 74.11kg