Quantum Approximate Optimization Algorithm (QAOA)
QAOA is a variationally-optimized sampling method used to obtain a probability distribution over the entire peptide conformational space.
Input Parameters
| Parameter | Constraint | Description |
|---|---|---|
| Sequence Length | < 8 Amino Acids | Strictly limited by QAOA block circuit depths. |
| P (Steps) | 1 – 5 Layers | Controls the number of alternating phase and mixing operators. |
| Lattice Constraints | 2D or 3D Grid | Enforced by XY mixers to guarantee valid physical geometries. |
Output Metrics
| Metric Name | Range | What it Means |
|---|---|---|
| Conformation Probability | 0 – 1.0 (Distribution) | The relative certainty of encountering a given physical geometry. |
| Shot Count | 1000 – 10000+ | Sample executions taken to compile stable density matrices. |
| Valid Folds | % of Shots | Ratio of shots that resulted in a physically un-collapsed structure. |
When to use QAOA
QAOA is highly practical when the task revolves around discovering probability spectra. Uses involve observing alternative local-minimum conformations generated by mutations. Use VQE if solely targeting the singular optimal ground state.
Benchmark Results
TARGET SEQUENCE: AGGAGA
Sampling via QAOA (p=2 block) on IBM Quantum returned a stable distribution yielding a 0.45 maximum probability corresponding to the lowest energy fold. Over 96% of all samples maintained strict self-avoiding physical constraints.
Source: Boulebnane et al. (2023)