Bimodal frequency distribution of size of protein loops

Question

The graph of number of amino acid (AA) residues in a loop Vs the frequency of their occurrence in proteins largely follows a tending-to-zero pattern. However, there appear to be some specific number of AAs where the frequency suddenly drops, leading to bi- or multi-modality in the plot. Is there some molecular explanation for this, or is it just an artifact?

I’ve attached the figure from a paper with a different research question but the same graph(Figure 2), albeit more detailed. (The same graph can be found in Brandon and Tooze (fig.2.8).) In the paper from which this is taken “A loop structure was defined as any region between two regular secondary structures that was at least three residues in length.”

Answer 1

A curve of this type, showing a bi- or multi-modal distribution, suggests two or more constituent species in the population. In this case this is borne out by further analysis in the paper of different types of secondary structure components anchoring the loop, the peak at low numbers of amino acids reflecting the loops connecting β-strands in anti-parallel β-sheets.

The second frame (B) of Figure 2 is shown below (dashed line):

Anti-parallel β-sheets are abundant in proteins, and in many — although not all — cases adjacent strands are connected by short (four- or five-amino acid) loops such as the β-turn, the β-hairpin or the β-bulge loop.

Repeating small loops are especially well illustrated in β-sheets displaying β-meander:

The β-meander motif from the outer surface Protein A

Examples in larger domains of the β-barrel and immunoglobulin type also show some larger loops.

Beta barrel of green fluorescent protein, showing loops at end of strands

Immunoglobulin domain of PDB 5IML showing β-sheet with loops connecting strands