# Finding conformations of peptide based on dihedral angles

I was trying to look at the alanine dipeptide trajectories from this link and track the trajectories in a Ramachandran plot. Below is a plot based on dihedral angles at each timestep for all trajectories (three simulations).

$$\qquad\qquad\qquad$$

I was wondering what is the criteria (based on dihedral angles) to classify each configuration based on it's conformation and how many classes of conformations do we have in total? Something like the following picture:

$$\qquad$$

• +1. It will be an excellent contribution to this site, to get the software recommendations for this task. The second question could even be a separate question ("does any software do this?" vs "how many conformations do we have in total?") – Nike Dattani Sep 2 '20 at 12:51
• I partially address this question in your other question about software for this purpose. Check out the paper there and see if it answers your question. If it does, I'm fine with you taking what you learned and writing your own answer here. – Tyberius Sep 2 '20 at 16:25
• Blade, have you considered Tyberius's last comment and perhaps learned enough to write an answer here? – Nike Dattani Sep 26 '20 at 21:04

I don't think that any classification of dihedral angles exists.

α and β that you see in the picture refer to alpha helices and beta sheets – structural elements of proteins (secondary structure). What is neither a helix nor β-strand is called a loop or coil. Sometimes more detailed categories are used – for example, 310 and π helices can be used as separate categories.

So it's not that dipeptides are classified into multiple categories. It's that secondary structural elements of proteins have certain dihedral angles.

As an illustration, below is a figure from the Ramachandran's paper from 1963. The author marks there, for example, where the ideal left-handed α-helix is. But the same φ/ψ angles can also happen in a loop.