Alignment of multiple metabolomics LC-MS datasets from disparate diseases to reveal fever-associated metabolites.
Alignment of multiple metabolomics LC-MS datasets from disparate diseases to reveal fever-associated metabolites.
复制标题
DOI：
10.1371/journal.pntd.0011133
复制
发表时间：
2023-07
影响因子：
3.8
通讯作者：
中科院分区：
文献类型：
--
作者：
急性发热性疾病仍然是全球死亡率和发病率的主要原因，特别是在低收入和中等收入国家。这项研究的目的是确定感染了引起发烧的不同病原体的患者的任何可能的代谢共性。使用了三个液质联用(LC-MS)数据集，研究了疟疾、利什曼病和寨卡病毒感染对代谢的影响。数据集之间的保留时间(RT)漂移使用从LC-MS实验的质量控制中通常使用的内部标准获得的界标来确定。使用拟合的高斯过程模型(GP)对实验之间的RT漂移进行高水平校正，然后在样品与三个LC-MS数据集的校正RT之间进行标准峰集比对。随后对整合的峰集进行统计分析、注释和通径分析。确定了在所有数据集中常见的代谢失调模式，犬尿氨酸途径是所有三个与发热相关的数据集之间受影响最严重的途径。与发烧相关的传染病仍然是中低收入国家令人担忧的一个主要原因。对误诊疾病的不当治疗可能会导致耐药微生物的选择。因此，改进对发热患者的诊断和发现特定的生物标记物以支持新的诊断是可取的。代谢组学研究可以为生物标志物的发现提供必要的信息。在这项研究中，我们研究了三个不同的代谢组学数据集；包括疟疾、利什曼病和寨卡病毒感染的数据集，所有这些数据集都与发烧有关。我们的目标是整合这些代谢组学数据集，以确定在不同传染病中表现相同的代谢物。整合代谢组学数据集的挑战之一是它们之间在保留时间方面出现的非线性漂移。在这种情况下，我们建议使用一种称为高斯过程回归的有监督的机器学习算法来纠正这种漂移。在整合或对齐数据集之后，对代谢物进行统计分析和注释。我们确定了几种代谢物，这些代谢物在数据集中的作用方式相似，特别是那些在色氨酸代谢的犬尿氨酸途径中发现的代谢物。
Acute febrile illnesses are still a major cause of mortality and morbidity globally, particularly in low to middle income countries. The aim of this study was to determine any possible metabolic commonalities of patients infected with disparate pathogens that cause fever. Three liquid chromatography-mass spectrometry (LC-MS) datasets investigating the metabolic effects of malaria, leishmaniasis and Zika virus infection were used. The retention time (RT) drift between the datasets was determined using landmarks obtained from the internal standards generally used in the quality control of the LC-MS experiments. Fitted Gaussian Process models (GPs) were used to perform a high level correction of the RT drift between the experiments, which was followed by standard peakset alignment between the samples with corrected RTs of the three LC-MS datasets. Statistical analysis, annotation and pathway analysis of the integrated peaksets were subsequently performed. Metabolic dysregulation patterns common across the datasets were identified, with kynurenine pathway being the most affected pathway between all three fever-associated datasets. Fever-associated infectious diseases are still a major cause of concern in low to middle income countries. Inappropriate treatment of misdiagnosed diseases can contribute to the selection of drug resistant microbes. Therefore, improved diagnostics of febrile patients and specific biomarker discovery to support new diagnostics is desirable. Metabolomics studies can provide the necessary information which leads to the discovery of biomarkers. In this study we have investigated three different metabolomics datasets; including those for malaria, leishmaniasis and Zika virus infection, all associated with fever. We aimed to integrate these metabolomics datasets to determine metabolites which behave in the same way in different infectious diseases. One of the challenges in integrating metabolomics datasets is a non-linear drift which occurs between them in terms of retention time. In this case, we proposed to correct this drift by using a supervised machine learning algorithm called Gaussian Process Regression. Following the integration or alignment of the datasets statistical analysis and annotation of the metabolites was performed. We identified several metabolites which acted in a similar manner across the datasets, specifically those found in the kynurenine pathway of tryptophan metabolism.
登录
查看更多内容
DOI：
10.1038/nrm.2017.107
发表时间：
2018-03
期刊：
Nature reviews. Molecular cell biology
影响因子：
--
作者：
Hannun YA;Obeid LM
通讯作者：
Obeid LM
影响因子：
7.4
作者：
Domingo-Almenara X;Montenegro-Burke JR;Benton HP;Siuzdak G
通讯作者：
Siuzdak G
影响因子：
4.8
作者：
Igarashi, J;Bernier, SG;Michel, T
通讯作者：
Michel, T
影响因子：
7.4
作者：
Habra, Hani;Kachman, Maureen;Bullock, Kevin;Clish, Clary;Evans, Charles R.;Karnovsky, Alla
通讯作者：
Karnovsky, Alla
DOI：
10.1016/j.jasms.2003.09.011
发表时间：
2004-01-01
影响因子：
3.2
作者：
Lioe, H;O'Hair, RAJ;Reid, GE
通讯作者：
Reid, GE