峰值侦探:一种基于深度学习的半监督方法,用于在非目标代谢学中达到峰值治愈
Ethan Stancliffe1,2,3, Gary J Patti1,2,3,4
1Department of Chemistry, Washington University in St. Louis, St. Louis, Missouri 63130, United States.
Analytical chemistry
|June 14, 2023
概括
PeakDetective是一款新的深度学习工具,可以准确地区分真正的代谢物与非目标代谢学数据中的工件. 这种自动化方法显著改善了研究人员的数据质量和下游分析.
科学领域:
- 计算生物学 计算生物学
- 分析化学 分析化学
- 生物化学 生物化学
背景情况:
- 未定位的代谢学产生了大量文物的大型数据集,使分析复杂化.
- 当前的峰值检测算法优先考虑灵敏性而不是选择性,导致高错误阳性率.
- 现有的文物移除方法需要大量的手动干预,因为不同的峰形状.
研究的目的:
- 开发一种自动化,准确的方法来将峰值分类为真实代谢产物或代谢学数据中的工件.
- 为了减少数据处理中由丰富文物峰值列表引起的瓶.
- 提高非目标代谢学数据分析的可靠性和效率.
主要方法:
- 开发了PeakDetective,这是一个半监督的深度学习方法,用于峰值分类.
- 使用无监督的自动编码器将特征提取到隐藏空间.
- 使用主动学习来快速进行分类器培训,使用最小的用户标记数据.
主要成果:
- 与现有方法相比,PeakDetective在五个不同的LC/MS数据集中在删除文物方面表现出卓越的准确性.
- 积极学习方法需要最小的用户输入 (少于100个峰值) 快速模型培训.
- 对SARS-CoV-2数据集的应用确定了更多具有统计意义的代谢物,增强了生物见解.
结论:
- PeakDetective提供了一个高度准确和高效的解决方案,用于处理非目标代谢学数据.
- 该工具可以快速适应特定的分析方法和样本类型.
- PeakDetective 改进了代谢物识别,使生物发现更加稳健,并促进了传染病等领域的研究.
相关概念视频
¹³C NMR: Distortionless Enhancement by Polarization Transfer (DEPT)
When proton-coupled carbon-13 spectra are simplified by a broadband proton decoupling technique, structural information about the coupled protons is lost. Distortionless enhancement by polarization transfer (DEPT) is a technique that provides information on the number of hydrogens attached to each carbon in a molecule. While the DEPT experiment utilizes complex pulse sequences, the pulse delay and flip angle are specifically manipulated. The resulting signals have different phases depending on...
High-Performance Liquid Chromatography: Types of Detectors
The role of the detectors in High-Performance Liquid Chromatography (HPLC) is to analyze the solutes as they exit from the chromatographic column. The detector recognizes the solute's property and generates corresponding electrical signals, which are converted into a readable graph of the detector's response versus elution time called a chromatogram at the computer. There are several types of HPLC detectors, each with its own advantages and limitations, depending on the analyte properties and...


