基于图形对比学习和部分特征掩盖的分子性质预测.
Kunjie Dong1, Xiaohui Lin1, Yanhui Zhang1
1School of Computer Science and Technology, Dalian University of Technology, Dalian, 116024, China.
Journal of molecular graphics & modelling
|March 22, 2025
概括
这项研究引入了一种新的分子图对比学习方法 (FMGCL),以改善分子性质预测. FMGCL通过掩盖特征来增强样本生成,保留化学语义,以便更好地获得先前的知识.
科学领域:
- 计算化学是一种计算化学.
- 机器学习在药物发现中的作用
背景情况:
- 分子表示学习对于诸如分子性质预测 (MPP) 和药物设计等任务至关重要.
- 自主监督学习 (SSL),特别是对比学习 (CL),通过学习可概括的分子知识来解决MPP中的数据稀缺问题.
- 生成语义上保存的增强样本是CL对分子数据的一个关键挑战.
研究的目的:
- 提出一种新的对比学习框架,FMGCL,用于增强分子表示学习.
- 为了应对产生有效增强分子样本的挑战,这些样本保留了核心化学语义.
- 通过更好的预训练分子编码器,提高下游任务如MPP的性能.
主要方法:
- 开发了FMGCL,一个基于部分特征掩饰的图形对比学习模型.
- 通过掩盖部分原子和键特性来构建掩盖的分子图形,保持分子结构和语义.
- 集成批次内的相对样本距离,以提高回归任务的性能.
主要成果:
- 在来自MoleculeNet和ChEMBL的12个基准数据集上,FMGCL表现出卓越的性能.
- 拟议的部分特征掩盖策略在增强过程中有效地保留了分子语义.
- 该方法在预训练期间成功捕获了有价值的先前分子知识.
结论:
- FMGCL为分子表示学习提供了一个强大的方法,优于现有的方法.
- 在增强样本中保存化学语义对于化学信息学中有效的对比学习至关重要.
- FMGCL框架为在药物发现和分子性质预测方面推进机器学习应用提供了有希望的方向.
相关概念视频
Masking and Demasking Agents
2.3K
EDTA titrations may necessitate masking and demasking agents to temporarily protect a particular metal ion in a mixture from the EDTA reaction. These agents facilitate the sequential analysis of the metal ions by forming stable complexes with some—but not all—metal ions during certain steps.
There are many masking agents, such as cyanide, fluoride, triethanolamine, thiourea, and 2,3-bis(sulfanyl)propan-1-ol (formerly 2,3-dimercapto-1-propanol), with the masking agent chosen based on...
There are many masking agents, such as cyanide, fluoride, triethanolamine, thiourea, and 2,3-bis(sulfanyl)propan-1-ol (formerly 2,3-dimercapto-1-propanol), with the masking agent chosen based on...
2.3K
Predicting Molecular Geometry
34.0K
VSEPR Theory for Determination of Electron Pair Geometries
34.0K
Difference from Background: Limit of Detection
5.2K
The limit of detection (LOD) is the smallest amount of analyte that can be distinguished from the background noise. The LOD value corresponds to the concentration at which the analyte signal is three times larger than the standard deviation of the blank signal. Below this value, the analyte signal cannot be differentiated from the background noise. It is calculated by dividing the calibration slope by 3 times the standard deviation of the blank signals.
The LOD indicates the presence or absence...
The LOD indicates the presence or absence...
5.2K
Predicting Products: Substitution vs. Elimination
11.3K
When a nucleophile and an alkyl halide react, nucleophilic substitution and β-elimination reactions compete to generate products.
The following factors can influence the mechanisms competing against each other:
The following factors can influence the mechanisms competing against each other:
11.3K
Predicting Products: SN1 vs. SN2
13.1K
Nucleophilic substitution reactions of alkyl halides can proceed via an SN1 or an SN2 mechanism. While in SN2 reactions, the nucleophile attacks the substrate simultaneously as the leaving group departs, in SN1 reactions, the substrate first dissociates to give the carbocation intermediate. Various factors such as the structure of the substrate, the strength of the nucleophile, and the nature of the solvent promote one mechanism over the other.
With increased substitution on the alkyl halide,...
With increased substitution on the alkyl halide,...
13.1K
Molecular Models
37.7K
Physical models representing molecular architectures of chemical compounds play essential roles in understanding chemistry. The use of molecular models makes it easier to visualize the structures and shapes of atoms and molecules.
37.7K


