Testing the helix model for protein folding on four simple proteins

TL;DR

This study evaluates a simplified helix model's ability to predict protein secondary structures, finding it effective for synthetic proteins but less so for natural ones, highlighting the importance of non-local interactions.

Contribution

It introduces and tests a local helix model on multiple proteins, revealing the significance of non-local effects in accurate secondary structure prediction.

Findings

High accuracy on synthetic proteins

Lower accuracy on natural proteins

Non-local interactions are crucial for prediction

Abstract

We test a simplified, local version of the helix model on two synthetic and two natural proteins, to study its efficiency in predicting the native secondary structure. The results we obtain are very good for the synthetic sequences, poorer for the two natural ones. This suggests that non-local terms play a fundamental role in determining the secondary structure, even if in some cases local terms alone may be sufficient.