Simply put it, the type of analysis and the notation are seperate terms. You can apply each of big-O,little-o,big-Omga,little-omega, big theta and more (like tilde notations) to any of the analysis (including: best case, worst case, average case).
You have the type of analysis. This gives you the complexity function, like for example for merge sort's worst case:
T(n) = 2T(n/2) + CONST*n + SOME_CONSTANT
Then, you can analyse this T(n)
and derive some conclusions. The asymptotic notations are set of functions, or "family" of functions that have common asymptotic behavior. For the above examples, you can conclude:
- This function is in Theta(nlogn)
- This function is in O(nlogn)
- This function is in Omega(1)
- This function is in O(n^3)