Protein Structure Prediction with High Degrees of Freedom in a Gate-Based Quantum Computer

Area of Science:

• Computational Biology
• Quantum Computing
• Biophysics

Background:

• Protein structure prediction is vital for understanding biological functions.
• Current AI and simulation methods have limitations in accuracy and sampling for protein folding.
• Predicting 3D protein structure from amino acid sequences remains a significant challenge.

Purpose of the Study:

• To develop a novel quantum algorithm for protein structure prediction.
• To address the limitations of existing AI and classical simulation techniques.
• To accurately model the initial hydrophobic collapse in protein folding.

Main Methods:

• Developed a novel turn-based encoding algorithm for protein sequences.
• Utilized a gate-based quantum computer with up to 114 qubits (IBM hardware).
• Employed the simplified Hydrophobic-Polar (HP) model for protein representation.

Main Results:

• Successfully predicted protein structures of varied lengths using quantum computation.
• The algorithm demonstrated enhanced degrees of freedom compared to previous methods.
• The quantum formulation accurately captured the nucleation step, specifically the hydrophobic collapse.

Conclusions:

• The novel quantum algorithm offers a promising approach for accurate protein structure prediction.
• This method overcomes limitations of current AI and classical simulation techniques.
• The quantum computation successfully modeled the critical hydrophobic collapse in protein folding.