DeepPLM_mCNN: 基于预先训练的语言模型特征的多窗 CNN 增强离子通道和离子载体识别的方法
Van-The Le1, Muhammad-Shahid Malik2, Yi-Hsuan Tseng1
1Department of Computer Science and Engineering, Yuan Ze University, Chung-Li, 32003, Taiwan.
Computational biology and chemistry
|March 31, 2024
概括
DeepPLM_mCNN使用预训练语言模型 (PLM) 和多窗口卷积神经网络 (mCNN) 准确地分类膜蛋白,包括离子通道和输送器. 这种计算方法增强了蛋白质研究和药物开发.
科学领域:
- 计算生物学是一种计算生物学.
- 生物信息学是一种生物信息学.
- 分子生物学分子生物学
背景情况:
- 准确地分类膜蛋白对于理解细胞功能和开发新药至关重要.
- 离子通道和输送器是关键的膜蛋白类,在生理学和疾病中起着重要作用.
研究的目的:
- 开发一个新的计算框架,DeepPLM_mCNN,用于准确分类膜蛋白.
- 利用预训练语言模型 (PLM) 和多窗口卷积神经网络 (mCNN) 进行增强的特征提取和分类.
主要方法:
- 使用各种PLM (TAPE,ProtT5_XL_U50,ESM-1b,ESM-2_480,ESM-2_1280) 来从蛋白质序列中提取特征.
- 将PLM衍生特征集成到mCNN架构中,以识别对分类至关重要的保存图案.
- 对离子载体和离子通道的独立数据集进行模型性能评估.
主要成果:
- 基于ProtT5的模型实现了90%的灵敏度,95.8%的特异性和95.4%的准确性用于离子载体的分类.
- 基于ESM-1b的模型实现了88.3%的灵敏度,95.7%的特异性和95.2%的离子通道分类准确度.
- 在未见测试数据上,与现有方法相比,表现出显著的性能改进.
结论:
- 该DeepPLM_mCNN框架有效地使用序列数据单独对膜蛋白进行分类.
- 结合PLM和深度学习,为膜蛋白的计算识别提供了一个强大的方法.
- 这些发现支持了膜蛋白研究的进展,以及针对离子通道和输送器的向药物开发.
相关概念视频
Multi-pass Transmembrane Proteins and β-barrels
5.3K
In multi-pass transmembrane proteins, the polypeptide chain crosses the membrane more than once. The transmembrane polypeptide chain either forms an α-helix or β-strand structure. α-Helix containing multi-pass transmembrane proteins are ubiquitous, whereas β-strand containing ones are mainly found in gram-negative bacteria, mitochondria, and chloroplasts.
α-Helix containing multi-pass transmembrane proteins
Multi-pass transmembrane proteins such as...
α-Helix containing multi-pass transmembrane proteins
Multi-pass transmembrane proteins such as...
5.3K
Ligand-gated Ion Channels
12.4K
Ligand-gated ion channels are transmembrane proteins with a channel for ions to pass through and a binding site for a ligand. The channel opens only when a ligand attaches to the binding site.
Three Subfamilies of Ligand-gated Ion Channels
Ligand-gated ion channels fall into three subfamilies. The 'Cys-loop' includes the nicotinic acetylcholine receptors, γ-aminobutyric acid (GABA), glycine, and 5-hydroxytryptamine receptors. The second one is the 'Pore-loop' channels that...
Three Subfamilies of Ligand-gated Ion Channels
Ligand-gated ion channels fall into three subfamilies. The 'Cys-loop' includes the nicotinic acetylcholine receptors, γ-aminobutyric acid (GABA), glycine, and 5-hydroxytryptamine receptors. The second one is the 'Pore-loop' channels that...
12.4K
Long-term Potentiation
55.2K
Long-term potentiation, or LTP, is one of the ways by which synaptic plasticity—changes in the strength of chemical synapses—can occur in the brain. LTP is the process of synaptic strengthening that occurs over time between pre- and postsynaptic neuronal connections. The synaptic strengthening of LTP works in opposition to the synaptic weakening of long-term depression (LTD) and together are the main mechanisms that underlie learning and memory.
55.2K
Non-gated Ion Channels
6.8K
Ion channels are specialized proteins on the plasma membrane that allow charged ions to pass down their electrochemical gradient. Their main function is to maintain the membrane potential which is critical for cell viability. These channels are either gated or non-gated and can transport more than a thousand ions within milliseconds for the cellular event to occur.
Compared to the gated ion channels, the non-gated channels, also known as leakage or passive channels, have no gating mechanism....
Compared to the gated ion channels, the non-gated channels, also known as leakage or passive channels, have no gating mechanism....
6.8K


