Controlling the quality of metabolomics data: new strategies to get the best out of the QC sample

Abstract

The type and use of quality control (QC) samples is a 'hot topic' in metabolomics. QCs are not novel in analytical chemistry; however since the evolution of using QCs to control the quality of data in large scale metabolomics studies (first described in 2011), the need for detailed knowledge of how to use QCs and the effects they can have on data treatment is growing. A controlled experiment has been designed to illustrate the most advantageous uses of QCs in metabolomics experiments. For this, samples were formed from a pool of plasma whereby different metabolites were spiked into two groups in order to simulate biological biomarkers. Three different QCs were compared: QCs pooled from all samples, QCs pooled from each experimental group of samples separately and QCs provided by an external source (QC surrogate). On the experimentation of different data treatment strategies, it was revealed that QCs collected separately for groups offers the closest matrix to the samples and improves the statistical outcome, especially for biomarkers unique to one group. A novel quality assurance plus procedure has also been proposed that builds on previously published methods and has the ability to improve statistical results for QC pool. For this dataset, the best option to work with QC surrogate was to filter data based only on group presence. Finally, a novel use of recursive analysis is portrayed that allows the improvement of statistical analyses with respect to the ratio between true and false positives.

Missing value imputation strategies for metabolomics data

“…One common strategy, and the one employed in this research, is to filter variables based on exhibiting presence in at least 75% of samples in at least one of n groups. Missing values due to reprocessing errors can be reduced by methods from simply combining duplicate measurements in the peak picking process and taking an average of each [6] to applying a target ion search based on a predefined library such as the previously described recursive analysis [1]. For those missing values that still occur after taking actions such as these, they can be dealt with by changing data-reprocessing parameters or manual assignation of values from raw data [7,8], or imputed by zero [9], median [5], minimum value [10], ½ minimum value [11], arithmetic mean of all [11,12] or some of the more related samples [3], k-means nearest neighbor (kNN) [2,13] etc.…”

“…The molecular feature extraction tool in Mass Hunter Qualitative Analysis (B.06.00, Agilent) was used to clean data of background noise and to provide a list of all possible features in each sample (as described previously [1]). Features were created using the accuracy of mass measurements to group ions related to the charge-state envelope, isotopic distribution, and/or the presence of adducts and dimers, as well as potential neutral loss of molecules.…”

“…Alignment was performed by restricting the number of ions and charge states defined previously during the extraction of features. Finally, data were filtered based on a previously optimised strategy known as QA+ [1], whereby variables were retained if they were present in at least 50% of the QCs or absent in QCs, present in 100% across the two groups and present with a QC relative standard deviation below 30% after replacing one or two missing values out of eight with NaNs and leaving all others as zero. The adaptation of the original filtering method to accept only variables present in 100% of both groups was performed in order to aid the preparation of test sets as described in the next section.…”

“…Missing values in metabolomics data are reported to comprise around 10-40% of data [1]. Whether they should (because they really are absent) or should not affect the determination of significance between groups, it is suspected that missing data contributes to misinterpretation of metabolomics data and therefore this topic receives significant attention in the omics community.…”

Technical Challenges in Mass Spectrometry-Based Metabolomics

“…Recent large-scale metabolome analyses employed an experimental design using QC samples to correct a dri of the raw signal intensity during the analysis. [85][86][87] e QC samples were prepared by mixing all the sample extracts in one analysis batch or in one metabolome analysis study. e iterative analyses of the QC sample were inserted into the start, end, and between every 4-8 actual samples in batch sequences of data acquisition.…”