FusPB-ESM2:ProtBERT和ESM-2的融合模型用于细胞透预测
Fan Zhang1, Jinfeng Li1, Zhenguo Wen1
1Beijing Institute of Petrochemical Technology, Beijing 102617, China.
Computational biology and chemistry
|May 31, 2024
概括
这项研究引入了一种用于预测细胞透的新计算模型. 先进的功能融合方法实现了高精度,改善了药物输送和基因疗法开发.
科学领域:
- 生物化学 生物化学
- 计算生物学 计算生物学
- 药物运输 药物运输 药物运输
背景情况:
- 细胞透 (CPP) 对于提高药物的生物可用性和通过促进细胞膜传输来实现基因疗法至关重要.
- 目前的预测方法,包括传统的湿实验室技术和现有的计算方法,在速度,成本,准确性和可靠性方面存在局限性.
- 提高CPP预测的效率和准确性对于推进治疗应用至关重要.
研究的目的:
- 开发一个高度准确和可靠的计算模型来预测细胞透.
- 通过利用先进的特征提取和融合技术,克服现有预测方法的局限性.
- 为识别潜在的CPP提供一个具有成本效益和时间效率的替代方案.
主要方法:
- 使用蛋白质预训练语言模型,特别是ProtBERT和ESM-2,作为高级特征提取器.
- 实施了特征融合策略,将两种语言模型的表征结合起来,创建更全面的特征集.
- 采用线性映射用于基于融合特征的最终预测.
- 通过对公开可用的数据集进行广泛的实验,验证了模型的性能.
主要成果:
- 提出的基于特征融合的预测模型在识别细胞透上表现出高准确性和可靠性.
- 实现了0.983的高曲线下面积 (AUC) 值,表明优异的预测性能.
- 合并的特征表示对于CPP预测比单个模型特征更有效.
- 该模型比传统和现有的计算方法有显著的改进.
结论:
- 开发的计算模型代表了细胞透预测的重大进步.
- 使用ProtBERT和ESM-2等蛋白质预训练语言模型的特征融合提高了预测的准确性和可靠性.
- 这种方法提供了一个有希望的,高效的和具有成本效益的工具,可以加速治疗应用的CPP的发现和开发.
- 高AUC值强调了该模型对药物输送和基因治疗研究的影响潜力.
更多相关视频
相关概念视频
01:49Conserved Binding Sites
5.3K
Many proteins’ biological role depends on their interactions with their ligands, small molecules that bind to specific locations on the protein known as ligand-binding sites. Ligand-binding sites are often conserved among homologous proteins as these sites are critical for protein function.
Binding sites are often located in large pockets, and if their location on a protein’s surface is unknown, it can be predicted using various approaches. The energetic method computationally...
Binding sites are often located in large pockets, and if their location on a protein’s surface is unknown, it can be predicted using various approaches. The energetic method computationally...
5.3K
01:29Insertion of Multi-pass Transmembrane Proteins in the RER
19.7K
The rough ER membrane synthesizes, assembles, and embeds transmembrane proteins in diverse topologies. These proteins function as transporters or channels and can remain in the ER membrane or are sent to the Golgi complex, lysosome, and cell membrane.
The multipass transmembrane proteins are the type IV integral membrane proteins with multiple topogenic sequences determining their spatial arrangement in the ER membrane. Nearly all multipass proteins lack a cleavable signal sequence and use...
The multipass transmembrane proteins are the type IV integral membrane proteins with multiple topogenic sequences determining their spatial arrangement in the ER membrane. Nearly all multipass proteins lack a cleavable signal sequence and use...
19.7K