相关实验视频
Updated: Jul 17, 2025
10:25
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
Published on: November 11, 2022
8.9K
ProtEC:一个基于变压器的深度学习系统,用于准确地注释酶委员会号码
IEEE/ACM transactions on computational biology and bioinformatics
|September 4, 2023
概括
我们开发了一个基于变压器的深度学习模型,用于准确的酶注释. 该模型从蛋白质序列中预测酶委员会数,超过现有的机器学习方法.
科学领域:
- 生物信息学是一种生物信息学.
- 计算生物学 计算生物学
- 酶学 是一种酶学.
背景情况:
- 下一代测序已经创造了庞大的蛋白质数据库.
- 蛋白序列的手动注释是耗时和劳动密集的.
- 自动注释算法,包括深度学习,对于生物数据解释至关重要.
研究的目的:
- 开发一种基于变压器的新型深度学习模型,用于预测酶委员会 (EC) 数字.
- 为了从蛋白质序列中实现酶注释的最先进的准确性.
- 评估模型在不同序列数据集和不同训练大小中的稳定性.
主要方法:
- 采用了基于变压器的深度学习架构.
- 该模型经过训练,可以直接从完整的蛋白质序列中预测酶委员会号码.
- 在结构不同分布的聚类分割数据集上评估了性能.
主要成果:
- 拟议的模型在预测EC数字方面实现了最先进的准确性.
- 该模型在结构不同的数据集上表现出强的性能,表明深度模式识别.
- 精度在减少训练数据的情况下保持不变,并且独立于序列长度.
结论:
- 基于变压器的模型为自动化酶注释提供了一个高度准确和强大的解决方案.
- 它在序列变异中概括的能力使其适合于各种生物医学应用.
- 这种方法显著提高了大规模蛋白质序列数据库的实用性.
相关概念视频
01:00Enzyme-linked Receptors
78.7K
Enzyme-linked receptors are proteins that act as both receptor and enzyme, activating multiple intracellular signals. This is a large group of receptors that include the receptor tyrosine kinase (RTK) family. Many growth factors and hormones bind to and activate the RTKs.
Neurotrophin (NT) receptors are a family of RTKs, including trkA, trkB, and trkC (tropomyosin-related kinase) receptors. TrkA is specific for nerve growth factor (NGF), neurotrophin-6, and neurotrophin-7. TrkB binds...
Neurotrophin (NT) receptors are a family of RTKs, including trkA, trkB, and trkC (tropomyosin-related kinase) receptors. TrkA is specific for nerve growth factor (NGF), neurotrophin-6, and neurotrophin-7. TrkB binds...
78.7K
01:27Transducer Mechanism: Enzyme-Linked Receptors
2.5K
Enzyme-linked receptors are cell-surface receptors acting as an enzyme or associating with an enzyme intracellularly. They make excellent drug targets. Drugs can bind to the extracellular ligand-binding domain or directly affect their enzymatic domain and alter their activity.
Major types that are helpful drug targets include:
Major types that are helpful drug targets include:
2.5K