Paper List
-
Ill-Conditioning in Dictionary-Based Dynamic-Equation Learning: A Systems Biology Case Study
This paper addresses the critical challenge of numerical ill-conditioning and multicollinearity in library-based sparse regression methods (e.g., SIND...
-
Hybrid eTFCE–GRF: Exact Cluster-Size Retrieval with Analytical pp-Values for Voxel-Based Morphometry
This paper addresses the computational bottleneck in voxel-based neuroimaging analysis by providing a method that delivers exact cluster-size retrieva...
-
abx_amr_simulator: A simulation environment for antibiotic prescribing policy optimization under antimicrobial resistance
This paper addresses the critical challenge of quantitatively evaluating antibiotic prescribing policies under realistic uncertainty and partial obser...
-
PesTwin: a biology-informed Digital Twin for enabling precision farming
This paper addresses the critical bottleneck in precision agriculture: the inability to accurately forecast pest outbreaks in real-time, leading to su...
-
Equivariant Asynchronous Diffusion: An Adaptive Denoising Schedule for Accelerated Molecular Conformation Generation
This paper addresses the core challenge of generating physically plausible 3D molecular structures by bridging the gap between autoregressive methods ...
-
Omics Data Discovery Agents
This paper addresses the core challenge of making published omics data computationally reusable by automating the extraction, quantification, and inte...
-
Single-cell directional sensing at ultra-low chemoattractant concentrations from extreme first-passage events
This work addresses the core challenge of how a cell can rapidly and accurately determine the direction of a chemoattractant source when the signal is...
-
SDSR: A Spectral Divide-and-Conquer Approach for Species Tree Reconstruction
This paper addresses the computational bottleneck in reconstructing species trees from thousands of species and multiple genes by introducing a scalab...
用于量子退火优化的二元潜在蛋白质适应度景观
University of Alabama at Birmingham
30秒速读
IN SHORT: 通过将序列映射到二元潜在空间进行基于QUBO的适应度优化，桥接蛋白质表示学习和组合优化。
核心创新
- Methodology First framework to transform protein language model embeddings into binary latent representations for QUBO-based fitness modeling
- Methodology Enables direct compatibility with quantum annealing hardware through native QUBO formulation
- Biology Demonstrates that simple binary representations can capture meaningful structure in protein fitness landscapes
主要结论
- Q-BioLat在ProteinGym GFP数据集（10,000个样本，潜在维度32-64）上实现了0.385-0.413的Spearman相关性
- 优化后的序列始终检索到适应度百分位顶部的最近邻，模拟退火在代理分数上实现了1.529±的改进
- 遗传算法在更高维潜在空间（m=64）中优于其他方法，而局部搜索能更好地保持序列真实性
摘要: 我们提出了Q-BioLat，一个在二元潜在空间中建模和优化蛋白质适应度景观的框架。从蛋白质序列出发，我们利用预训练的蛋白质语言模型获得连续嵌入，然后将其转换为紧凑的二元潜在表示。在这个空间中，蛋白质适应度使用二次无约束二元优化（QUBO）模型进行近似，从而通过经典启发式方法（如模拟退火和遗传算法）实现高效的组合搜索。在ProteinGym基准测试中，我们证明Q-BioLat能够捕捉蛋白质适应度景观中的有意义结构，并能够识别高适应度变体。尽管使用了简单的二值化方案，我们的方法始终能检索到其最近邻位于训练适应度分布顶部的序列，特别是在最强配置下。我们进一步表明，不同的优化策略表现出不同的行为，进化搜索在更高维的潜在空间中表现更好，而局部搜索在保持真实序列方面仍具有竞争力。除了其经验性能外，Q-BioLat为蛋白质表示学习和组合优化之间提供了自然的桥梁。通过将蛋白质适应度表述为QUBO问题，我们的框架与新兴的量子退火硬件直接兼容，为量子辅助蛋白质工程开辟了新的方向。