Embarking on metabolomics analysis requires strategic decisions in LC-MS setup and meticulous data preprocessing. From choosing chromatographic columns to navigating high-resolution mass spectrometry and optimizing peak extraction, each step plays a crucial role.
LC-MS Configuration
The selection of LC-MS must address the biological questions posed and align with sample collection and preprocessing. To achieve comprehensive metabolite detection, diverse methods are considered in LC-MS analysis, including the use of different chromatographic columns.
LC
Reverse-phase chromatography, known for its wide applicability and stability, is extensively used in untargeted metabolomics. Commonly employed reverse-phase columns include traditional C18 and HSS T3 produced by Waters. C18 columns exhibit effective separation for most non-polar compounds like lipids and cholesterol. HSS T3 columns, with a low-density octadecyl ligand, enhance retention for polar compounds, broadening the range of detected compounds compared to conventional C18.
HILIC columns are suitable for analyzing highly polar metabolites, employing a higher proportion of organic phase in the mobile phase to improve ESI ionization and detection sensitivity. However, HILIC analysis exhibits lower reproducibility in retention time compared to reverse-phase chromatography. It is more sensitive to matrix effects, and various interactions between analytes and the column change with gradient variations, making it challenging to predict compound retention and elution sequences. Commonly used HILIC columns include amid-grafted silica gel (for polar compounds) and zwitterionic sulfobetaine stationary phase (for highly polar and ion-type compounds). Given the complexity of HILIC analysis, researchers without LC experience are advised to use established methods from literature or choose widely used column fillers, such as C18 in reverse-phase chromatography or BEH amide and ZIC-pHILIC in HILIC analysis.
Select Services
MS
High-resolution mass spectrometers can provide accurate molecular weights, isotope distributions, and tandem mass spectrometry (MS/MS) information, enhancing the precision of metabolite annotation and identification. Additionally, the use of ion mobility spectrometry to determine Collision Cross Section (CCS) information is increasingly applied in metabolomics studies. However, simultaneous acquisition of MS, MS/MS, and CCS information in a single analysis may lead to reduced instrument acquisition rates, sensitivity, and dynamic range, necessitating consideration in metabolomics analysis.
In untargeted metabolomics analysis, the two most commonly used types of high-resolution mass spectrometry are Orbitrap and Time-of-Flight (TOF) mass spectrometers. Both instrument types maintain a mass accuracy within 2 ppm. Q-TOF mass spectrometry achieves a resolution of 30,000 to 60,000 Full Width at Half Maximum (FWHM), sufficient for the analysis of small molecule metabolites. Orbitrap mass spectrometry provides higher resolution, ranging from 200,000 to 1,000,000 FWHM, but at the expense of duty cycle. TOF mass spectrometry, while maintaining high resolution, can achieve a maximum of 100 spectra per second, making Q-TOF mass spectrometry more suitable for the analysis of narrow LC chromatographic peaks.
High-resolution mass spectrometers can offer various scan modes for acquiring MS/MS information based on hardware and software configurations. One commonly used mode is Data-Dependent Acquisition (DDA), where ions exceeding intensity thresholds are selected for fragmentation. DDA provides clean MS/MS spectra and facilitates the establishment of relationships between precursor and fragment ions. However, a limitation of DDA is the inability to acquire fragment information for all compounds. Another approach is Data-Independent Acquisition (DIA), which does not preselect precursor ions for fragmentation, theoretically allowing the fragmentation of all ions. However, DIA produces mixed fragment ion spectra for all precursor ions, making it challenging to establish relationships between precursor and fragment ions. In recent years, the SWATH method has emerged, dividing the precursor ion scan range into continuous isolation windows. This approach reduces the number of precursor ions for fragment ion acquisition, simplifying the complexity of fragment spectra, and enabling deconvolution to obtain precursor-to-fragment ion relationships. The incorporation of ion mobility technology significantly increases peak capacity in high-resolution mass spectrometry detection and effectively improves the quality of fragment spectra obtained in DIA. The CCS values provided by ion mobility technology also offer reliable information for compound identification.
Learn more
An Optimised MS-Based Versatile Untargeted Metabolomics (Marques et al., 2023)
Peak Extraction
Peak extraction involves striking a balance between retaining most of the useful mass spectrometry information and discarding unrelated data, transforming LC-MS data into a structured dataset. The raw data from LC-MS is a single file containing continuous time, mass, and intensity information. A series of files needs to be converted into a dataset containing retention time, mass-to-charge ratio, and peak intensity, a step known as peak picking. The raw data initially undergoes filtering to eliminate random noise. Subsequently, an intensity threshold is defined to select peaks for extraction, and the definition of this threshold requires optimization to avoid losing key compounds with lower intensity or retaining excessive noise. Both commercial and open-source software can be employed for this step, such as Waters' Progenesis QI, open-source tools like XCMS, mzMine, and MS-DIAL, among others.
Since retention time drift occurs with the run of the analysis sequence, alignment algorithms are employed to adjust the retention time between the same analysis sequences or even different batches of samples, ensuring that peaks of the same ion have consistent retention times across all samples. Prior to peak extraction, alignment of retention times is necessary to ensure accurate mass-to-charge ratio and retention time for a chromatographic peak in all samples, facilitating measurement.
In LC-MS analysis, a compound can produce different ions, such as various adduct forms or in-source fragmentation ions, leading to dataset complexity. Peak grouping or deconvolution operations can group different ions belonging to the same compound based on retention time matching and mass-to-charge ratio differences between different peaks (adduct ions or neutral losses), streamlining the dataset.
Data Correction and Scaling
Quality control (QC) samples span the entire experimental analysis, possessing identical compound compositions with a single source of signal variation, making them suitable as a baseline for data correction. The main aspects of correction include:
1. Filtering: Use continuous injection QC and dilution QC samples to filter data, ensuring that peaks of interest originate from the samples themselves. The response intensity of extracted peaks should correlate with their dilution factors.
2. Drift: Due to uncontrollable factors, the signal response of compounds at a fixed concentration may vary over time. As the concentration of compounds in QC samples remains constant, assessing the response of compounds over time in these samples allows evaluation of variations in compound response throughout the entire sequence run.
3. Data Scaling: In untargeted metabolomics analysis, obtaining the true concentration of each compound from mass spectrometry detection is challenging. For instance, a compound with a low concentration may produce a higher response signal due to easy ionization, while a high-concentration compound may have a lower signal intensity. Therefore, high-intensity compounds in the analysis results may not necessarily have high concentrations in the samples. Similarly, compounds with the greatest intensity changes may not exhibit significant concentration differences.
Reference
- Marques, Cátia F., and Gonçalo C. Justino. "An Optimised MS-Based Versatile Untargeted Metabolomics Protocol." Separations 10.5 (2023): 314.
You may be interested in:
