It seems just obvious that it's at least a sampling problem. Assuming an average protein length of 400 amino acids and 20 possible amino acids, that's about 10^520 different possibilities for sequences, which is a mind-bogglingly large number.

We haven't even begun to explore the biological universe.