Helical structures are prevalent in biology. In the PDB, there are many examples where protein molecules are helically arranged, not only according to strict crystallographic screw axes but also according to approximate noncrystallographic screws. The preponderance of such screws is rather striking as helical arrangements in crystals must preserve an integer number of subunits per turn, while intuition and simple packing arguments would seem to favor fractional helices. The article provides insights into such questions, based on stereochemistry, trigonometry, and topology, and illustrates the findings with concrete PDB structures. Updated statistics of Sohncke space groups in the PDB are also presented.