从SMILES表现中预测小分子的染色学保留时间,使用混合变压器-LSTM模型
Sargol Mazraedoost1, Hadi Sedigh Malekroodi2, Petar Žuvela1
1Department of Chemical Engineering, Pukyong National University, Busan 48513, Republic of Korea.
Journal of chemical information and modeling
|March 28, 2025
概括
这项研究引入了一种新的变压器-LSTM模型,用于使用分子结构预测液体色谱保留时间. 该方法实现了高精度,优于现有的小分子分析方法.
科学领域:
- 计算化学计算化学
- 分析化学 分析化学
- 机器学习在化学中的应用
背景情况:
- 准确的保留时间 (RT) 预测对于液态色谱中的分子分析至关重要.
- 现有的RT预测方法在准确性和范围方面存在局限性.
- 需要新的计算方法来增强RT预测能力.
研究的目的:
- 开发和评估一种基于变压器的语言模型,用于预测反相高性能液体色谱 (RP-HPLC) 中的保留时间.
- 使用简化分子输入线输入系统 (SMILES) 序列作为RT预测的文本输入.
- 评估模型的性能,可解释性和转移学习能力.
主要方法:
- 采用了混合架构,将预训练的RoBERTa (强大优化的BERT方法) 与双向长期短期存储器 (BiLSTM) 网络相结合.
- 含有77,980个小分子的METLIN小分子保留时间 (SMRT) 数据集使用SMILES符号编码.
- 分子数据通过tokenizer处理以进行序列表示,并将多个变压器层的层融合纳入.
主要成果:
- 变压器-LSTM模型实现了卓越的性能,平均绝对误差 (MAE) 为26.23秒,平均绝对百分比误差 (MAPE) 为3.25%,R平方 (R2) 值为0.91.
- 注意力可视化表明模型能够识别影响RT的关键分子特征,表明可解释性.
- 该模型在PredRet数据库中的十个数据集中展示了强大的转移学习能力,显示了与先前方法相比的持续改进.
结论:
- 混合变压器-LSTM模型为预测液体色谱保留时间提供了有价值和准确的方法.
- 这种方法显示出在代谢学和更广泛的小分子分析中应用的巨大潜力.
- 使用SMILES作为变压器模型的文本输入,代表了染色体分析中的新有效策略.
更多相关视频
10:14Chromatographic Fingerprinting by Template Matching for Data Collected by Comprehensive Two-Dimensional Gas Chromatography
Published on: September 2, 2020
4.7K
08:56Detection of Regulated Ergot Alkaloids in Food Matrices by Liquid Chromatography-Trapped Ion Mobility Spectrometry-Time-of-Flight Mass Spectrometry
Published on: November 22, 2024
495
相关概念视频
Chromatographic Methods: Terminology
640
Chromatography is an analytical technique widely used in fields such as chemistry, biology, environmental science, and pharmaceuticals to separate the components of a mixture and identify substances between them. The process of chromatography is based on the interactions between two distinct phases: the stationary phase and the mobile phase. The stationary phase is fixed in place by a supporting material, while the mobile phase moves over it, carrying the solutes. As the mobile phase travels,...
640
Thin-Layer Chromatography (TLC): Overview
853
Thin-layer chromatography (TLC) is a chromatography technique that separates compounds based on their polarity. TLC typically uses polar silica gel, a form of silicon dioxide, as the stationary phase. The silica gel contains hydroxyl (OH) groups on its surface, which form hydrogen bonds with polar compounds, influencing their adhesion to the stationary phase.
To begin the analysis, a mixture of compounds is spotted on the starting line on the TLC plate using a thin capillary. The bottom of the...
To begin the analysis, a mixture of compounds is spotted on the starting line on the TLC plate using a thin capillary. The bottom of the...
853
Silica Gel Column Chromatography: Overview
815
Silica gel column chromatography is a technique for separating compounds using a column packed with silica gel as the stationary phase. This method relies on differences in the polarity of compounds. Based on their polarities, compounds move between the stationary phase (silica gel) and the mobile phase (the solvent), forming discrete bands in the column.
Polar components tend to bind strongly to the silica gel, causing them to move slowly through the column. In contrast, nonpolar compounds...
Polar components tend to bind strongly to the silica gel, causing them to move slowly through the column. In contrast, nonpolar compounds...
815
Chromatographic Resolution
291
In chromatography, a solute moves through a chromatographic column and tends to spread, forming a Gaussian-shaped band. The longer the solute spends in the column, the broader the band becomes. The broadening can lead to overlaps within the column, affecting separation effectiveness.
The effectiveness of separation can be evaluated by determining the level of separation between two neighboring peaks in a chromatogram, which represents the individual components of a sample.
In chromatography,...
The effectiveness of separation can be evaluated by determining the level of separation between two neighboring peaks in a chromatogram, which represents the individual components of a sample.
In chromatography,...
291
