Home
This Title All WIREs
WIREs RSS Feed
How to cite this WIREs title:
WIREs Comput Mol Sci
Impact Factor: 8.836

Protein design: from computer models to artificial intelligence

Full article on Wiley Online Library:   HTML PDF

Can't access this content? Tell your librarian.

The rational design of novel biomolecules with desired functional properties is one of the most fascinating challenges in science, with implications at the fundamental and practical levels. From the fundamental point of view, the design of proteins able to support nonnatural reactivities represents the decisive test on our understanding of the molecular mechanisms through which biomolecules operate. From the practical point of view, new designs may open the way to applications in a wide variety of fields, ranging from health to life science, and from catalysis to material sciences. During the past decades, we have witnessed an amazing transition in the application of computational methods to protein and enzyme design, from simple fold prediction to ab initio structural design. Herein, we review key areas and fundamental aspects of research in the design of protein structures, interactions, and reactivities. We also provide our perspective on the exciting range of developments that are made possible by the integration of innovations in hardware, software, and theory, while keeping an eye on the applications, challenges, and opportunities that can open up in many different domains of science. WIREs Comput Mol Sci 2017, 7:e1318. doi: 10.1002/wcms.1318

This article is categorized under:

  • Structure and Mechanism > Computational Biochemistry and Biophysics
  • Molecular and Statistical Mechanics > Molecular Dynamics and Monte-Carlo Methods
A scheme for the integration of bioinformatics, structure‐based, and new machine learning and artificial intelligence methods in the design of new proteins with desired functions.
[ Normal View | Magnified View ]
Idealized model of an active site fit in the context of a three‐dimensional enzyme structure. Specifically, the transition state was computed by (SCC‐DFTB)/MM(AMBER) umbrella sampling molecular dynamics starting from the X‐ray structure of Rhodococcus erythropolis limonene epoxide hydrolase, PDB code no. 1nu3.
[ Normal View | Magnified View ]
A simplified representation of the SCHEMA approach, involving the recombination of stable substructures from different proteins of the same family. A two‐step learning algorithm is applied to the SCHEMA library to identify sequences that are both informative and functional.
[ Normal View | Magnified View ]
Directed evolution of enzyme sequences: sequences are selected on the fitness landscape to provide the final molecule with the desired functional properties.
[ Normal View | Magnified View ]
Simple representative scheme for grafting procedures. Starting from a structures database all the scaffold candidates are compared to the epitope via energetic, alignment, and packing scoring. The epitope is then transplanted into the best scoring scaffold. Finally, the grafted structure is further refined to improve stability.
[ Normal View | Magnified View ]
Different protein surfaces can recognize different binding partners: design can be aimed at the preferential recognition of a specific binder or at the promiscuous binding of multiple molecules, with different impact at the functional and cellular levels. Ran multispecific interface is shown: protein Ran (PDB access no. 1a2k) is given in green cartoons; Ran interface residues are displayed in sticks and color‐coded according to correspondent binding partners (ghost cartoons). Structurally characterized interaction partners refer to data in Ref .
[ Normal View | Magnified View ]

Browse by Topic

Molecular and Statistical Mechanics > Molecular Dynamics and Monte-Carlo Methods
Structure and Mechanism > Computational Biochemistry and Biophysics

Access to this WIREs title is by subscription only.

Recommend to Your
Librarian Now!

The latest WIREs articles in your inbox

Sign Up for Article Alerts