Introduction The generic metabolomics data processing workflow is constructed with a serial set of processes including peak picking, quality assurance, normalisation, missing value imputation, transformation and scaling. The combination of these processes should present the experimental data in an appropriate structure so to identify the biological changes in a valid and robust manner. Objectives Currently, different researchers apply different data processing methods and no assessment of the permutations applied to UHPLC-MS datasets has been published. Here we wish to define the most appropriate data processing workflow. Methods We assess the influence of normalisation, missing value imputation, transformation and scaling methods on univariate and multivariate analysis of UHPLC-MS datasets acquired for different mammalian samples.Results Our studies have shown that once data are filtered, missing values are not correlated with m/z, retention time or response. Following an exhaustive evaluation, we recommend PQN normalisation with no missing value imputation and no transformation or scaling for univariate analysis. For PCA we recommend applying PQN normalisation with Random Forest missing value imputation, glog transformation and no scaling method. For PLS-DA we recommend PQN normalisation, KNN as the missing value imputation method, generalised logarithm transformation and no scaling. These recommendations are based on searching for the biologically important metabolite features independent of their measured abundance. Conclusion The appropriate choice of normalisation, missing value imputation, transformation and scaling methods differs depending on the data analysis method and the choice of method is essential to maximise the biological derivations from UHPLC-MS datasets.

No abstract

Context:5α-Reductase 1 and 2 (SRD5A1, SRD5A2) inactivate cortisol to 5α-dihydrocortisol in addition to their role in the generation of DHT. Dutasteride (dual SRD5A1 and SRD5A2 inhibitor) and finasteride (selective SRD5A2 inhibitor) are commonly prescribed, but their potential metabolic effects have only recently been identified.Objective:Our objective was to provide a detailed assessment of the metabolic effects of SRD5A inhibition and in particular the impact on hepatic lipid metabolism.Design:We conducted a randomized study in 12 healthy male volunteers with detailed metabolic phenotyping performed before and after a 3-week treatment with finasteride (5 mg od) or dutasteride (0.5 mg od). Hepatic magnetic resonance spectroscopy (MRS) and two-step hyperinsulinemic euglycemic clamps incorporating stable isotopes with concomitant adipose tissue microdialysis were used to evaluate carbohydrate and lipid flux. Analysis of the serum metabolome was performed using ultra-HPLC-mass spectrometry.Setting:The study was performed in the Wellcome Trust Clinical Research Facility, Queen Elizabeth Hospital, Birmingham, United Kingdom.Main Outcome Measure:Incorporation of hepatic lipid was measured with MRS.Results:Dutasteride, not finasteride, increased hepatic insulin resistance. Intrahepatic lipid increased on MRS after dutasteride treatment and was associated with increased rates of de novo lipogenesis. Adipose tissue lipid mobilization was decreased by dutasteride. Analysis of the serum metabolome demonstrated that in the fasted state, dutasteride had a significant effect on lipid metabolism.Conclusions:Dual-SRD5A inhibition with dutasteride is associated with increased intrahepatic lipid accumulation.

Tandem mass spectrometry (MS/MS or MS) is a widely used approach for structural annotation and identification of metabolites in complex biological samples. The importance of assessing the contribution of the precursor ion within an isolation window for MS experiments has been previously detailed in proteomics, where precursor ion purity influences the quality and accuracy of matching to mass spectral libraries, but to date, there has been little attention to this data-processing technique in metabolomics. Here, we present msPurity, a vendor-independent R package for liquid chromatography (LC) and direct infusion (DI) MS that calculates a simple metric to describe the contribution of the selected precursor. The precursor purity metric is calculated as "intensity of a selected precursor divided by the summed intensity of the isolation window". The metric is interpolated at the recorded point of MS acquisition using bordering full-scan spectra. Isotopic peaks of the selected precursor can be removed, and low abundance peaks that are believed to have limited contribution to the resulting MS spectra are removed. Additionally, the isolation efficiency of the mass spectrometer can be taken into account. The package was applied to Data Dependent Acquisition (DDA)-based MS metabolomics data sets derived from three metabolomics data repositories. For the 10 LC-MS DDA data sets with > ±1 Da isolation windows, the median precursor purity score ranged from 0.67 to 0.96 (scale = 0 to +1). The R package was also used to assess precursor purity of theoretical isolation windows from LC-MS data sets of differing sample types. The theoretical isolation windows being the same width used for an anticipated DDA experiment (±0.5 Da). The most complex sample had a median precursor purity score of 0.46 for the 64,498 XCMS determined features, in comparison to the less spectrally complex sample that had a purity score of 0.66 for 5071 XCMS features. It has been previously reported in proteomics that a purity score of <0.5 can produce unreliable spectra matching results. With this assumption, we show that for complex samples there will be a large number of metabolites where traditional DDA approaches will struggle to provide reliable annotations or accurate matches to mass spectral libraries.

No abstract