Lipidomics and Metabolomics services
In the context of mass spectrometry, Lipidomics and Metabolomics are defined as the study of lipid and small molecule metabolites, respectively, through the mass spectrometric detection and quantification of one or more of their ionic forms. Samples are often complex mixtures of polar and non-polar metabolites, and the type of biological question behind the analysis dictates the type of experiment to be conducted. A user may be interested in the quantification of specific analytes (targeted acquisition) or detect as many analytes as possible without prior knowledge (untargeted acquisition). Different experiments require different amount of starting material and sample preparation protocols. The next section will discuss the type of experiments offered by the LSU MSF and the associated details. The LSU MSF utilizes a Waters platform comprising of a Synapt XS Q-TOF with ion mobility and an Acquity Premiere UPLC system.
MSE refers to a data independent acquisition mode that is performed on Waters instruments, where the acquisition is divided into two functions, with data point collected alternating between them. The first function (the low energy function) is a survey scan of the mass range specified for the analysis, with the collision energy set to a low level that produces minimal (if any) fragmentation. The second function (the high energy function) is also acquired using the entire mass range, but with the fragmentation energy turned on (or ramped, depending on the experiment). This produces fragment ions from all parent ions present in the previous scan. After the acquisition is completed, the software evaluates the two functions and aligns parent ions and fragments based on their chromatographic profiles, generating reconstructed MS/MS spectra. When ion mobility is present, the system operates in HDMSE mode and records mobiligrams of the acquired ions in addition to spectra. The mobiligrams for the high energy function is acquired before the fragmentation, thus allowing the use of the ion mobility function to further sort which fragments belong to which parent ion by using a similar alignment algorithm to that used for the chromatographic alignment of peaks in MSE. The result is cleaner reconstructed MS/MS spectra, although some ion type do not tolerate well the ion mobility separation.
Description: this type of analysis provides a general overview of both porla and non-polar components of a sample. While not targeting specific compounds, the method is able to detect a broad number of molecular classes, from amino acids to TCA cycle molecules, from non-polar hormones to fatty acids. At the same time, the method does not have the same sensitivity as more targeted approaches that aim to look at narrower windows of compound types. It is indicated for projects where there is little preliminary information (e.g., a plant that has not been analyzed before), or the experiment is testing a new condition (a new drug, or a new environmental factor), and there is a need to produce preliminary data, or funds are limited.
Amount needed: for tissue, as little as 1 mg, although 10 mg is the preferred amount. For cell samples, a minimum of 1x106 is preferred, but it is possible to obtain results from fewer cells as well.
Type of samples: Cell, tissue, organoids, biofilms etc.
How many replicates: the minimum is 3, but it is advisable to increase this number to 4 or 6 to take full advantage of the multivariate analysis that is often performed on the dataset.
MSF protocol available: yes.
Degree of difficulty: low.
Gradient length: generally between 10 and 15 minutes.
Notes: this method is ideal for quick screening projects. It allows the gathering of general information about what might be different between the conditions tested, or if there is a difference at all. For example, a user might be interested in measuring the metabolic effects of a new drug or checking if a knockout experiment is producing the type of metabolic changes that one might expect.
Description: this type of analysis specifically targets lipids, although it does not focus on one specific class. The analysis is performed on a charged surface hybrid (CSH) C18 column with solvent that includes ammonium formate, formic acid, acetonitrile and isopropyl alcohol specifically designed to maximize elution performance and detectability of as many lipids as possible. The mass spectrometer is operated in HMSE mode, which uses Waters “all ions fragmentation” approach coupled to ion mobility to increase the number of detected and identified ions.
Amount needed: for tissue, as little as 1 mg, although 10 mg is the preferred amount. For cell samples, a minimum of 1x106 is preferred, but it is possible to obtain results from fewer cells as well.
Type of samples: Cell, tissue, organoids, biofilms etc.
How many replicates: the minimum is 3, but it is advisable to increase this number to 4 or 6 to take full advantage of the multivariate analysis that is often performed on the dataset.
MSF protocol available: yes.
Degree of difficulty: low.
Gradient length: generally 18 minutes.
Notes: a known drawback of lipidomics is the high dynamic range between lipid types in a give sample. For example, some types of adipose tissue contain so much triglycerides that it is hard to detect anything else without overloading the column or the mass spectrometer detector. Also, lipidomics experiments should always include both positive and negative ionization. This approach can be performed in parallael with untargeted metabolomics on the same sample.
Description: this type of analysis specifically targets polar metabolites, although it does not focus on one specific class. The analysis is performed on a HILIC Amide column with acetonitrile and water containing formic acid. This method is designed to maximize the separation and detectability of small compounds such as amino acids, sugars, tricarboxylic acids etc. The mass spectrometer is operated in MSE mode, which uses Waters “all ions fragmentation” approach without ion mobility.
Amount needed: for tissue, as little as 1 mg, although 10 mg is the preferred amount. For cell samples, a minimum of 1x106 is preferred, but it is possible to obtain results from fewer cells as well.
Type of samples: Cell, tissue, organoids, biofilms etc.
How many replicates: the minimum is 3, but it is advisable to increase this number to 4 or 6 to take full advantage of the multivariate analysis that is often performed on the dataset.
MSF protocol available: yes.
Degree of difficulty: low.
Gradient length: generally 10 minutes.
Notes: this method does not utilize ion mobility. We experimentally determined that in the case of small molecules, the advantages of ion mobilities are not as pronounced as in the case of lipidomics, and that the signal loss for very small molecules (such as amino acids) is quite higher, leading to loss of signal even at relatively high concentrations. This approach can be performed in parallael with untargeted lipidomics on the same sample.
Description: this type of analysis utilizes the TOFMRM mode, where the mass spectrometer isolates a mass window of few Daltons at the time based on a user provided list, and fragments it. The resulting ions are used to generate the output signal similarly to parallel reaction monitoring on a trapping instrument. This approach is designed to target one molecule at the time and requires optimization of both chromatography and fragmentation conditions to obtain the maximum sensitivity possible. Limits of detection in the part per billion range are generally achieved.
Amount needed: variable, depending on the type of sample and the target compound(s).
Type of samples: Cell, tissue, organoids, biofilms etc.
How many replicates: the minimum is 3, but the number is dictated by the biological variability of the target(s) under analysis.
MSF protocol available: yes.
Degree of difficulty: low.
Gradient length: generally 5-10 minutes.
Notes: these methods may or may not utilize ion mobility. The maximum acquisition rate for the Synapt XS is 30 Hz, which limits the number of transition that can be monitored when compared to dedicated instruments such as triple quadrupoles.
Untargeted data analysis is conducted on Waters Progenesis QI software (current version 3.0). Raw data are imported into the software and recalibrated using the accurate mass of the lockspray function. Runs are then aligned to each other to correct non-linear shifts in retention time and ensure that the same feature is measured across all runs. A pick peaking algorithm selects features that are then deconvoluted to group together ion types associated with the same feature (e.g., [M+H]+, M+Na]+, [2M+H]+ etc.). Deconvoluted features are then searched against databases using different approaches. The MSF utilizes an MS/MS database of experimentally measured spectra as primary database, when available. The primary MS/MS database for polar metabolites is METLIN 2019, while the primary MS/MS database for lipidomics is lipid blast. Each identification is associated with a score, which is the sum of 5 individual scores (fragmentation match, mass precision, isotopic pattern agreement, RT time agreement and collision cross section agreement), each contributing a maximum of 20 to the final. For most untargeted experiments retention time and collision cross section are either not known or not used, and therefore the maximum score is 60. We use a threshold of 40 to consider a match as a positive identification. If a feature has multiple identification with a score of 40 or more, the best scoring one is selected.
Features not matched using the MS/MS databases are submitted for another search using structure databases and theoretical fragmentation based on the following publication: Wolf, S., Schmidt, S., Müller-Hannemann, M. and Neumann, S., 2010. In silico fragmentation for computer assisted identification of metabolite mass spectra. BMC bioinformatics, 11, pp.1-12. Any SDF structure database can be used. For example, the MSF uses LipidMaps SDF files for lipidomics or the Human Metabolome Database (HMDB) for polar metabolites structure-based searches. Identified molecules are filtered based on the same threshold of 40 and assignments are based on the best scoring candidate.
The MSF is developing in-house curated databases that include MS/MS generated by our mass spectrometer and with our columns and gradients. This will increase the confidence of identification in untargeted experiments, but these efforts require a long time. The evolving databases are used as the first MS/MS database whenever possible.
Targeted analysis is performed with Waters MassLynx, which provides tools for visualization and integration of the acquired data. Further calculations are provided with Microsoft Excel.
Statistical analysis is mainly provided for untargeted experiments and makes use of both Waters Progenesis QI features as well as the Statistical package EZInfo, which is integrated with Progenesis QI. Possible analysis includes PCA, Power analysis, OPLS-DA and others.
The MSF is continuously developing an application intended for untargeted metabolomics and lipidomics, which is aimed at streamlining the data analysis that follows most (if not all) of the untargeted experiments we conduct: pathway analysis. There are several tools available in the literature, which offer several strategies to perform pathway analysis using the output of an LC-MS/MS experiment. At the same time, all strategies rely on R and Python code that requires moderate knowledge of computer coding in order to process the data and are often a hurdle for users. This project aims to create a simple interface that anyone with a Windows computer can use and allows utilization of these packages with no coding experience at all. The software is expected to be released in beta version in 2025.