The analytical variance was determined by measuring metabolite concentrations of a reference sample of pooled human plasma with four to six replicates on each of the six plates. All statistical analyses were performed with the statistic platform R, version 2. The median within-plate CV was 9.
- Attacks on Digital Watermarks Classification, Estimation-based Attacks and Benchmarks.
- Introduction | SpringerLink.
Reliability was expressed by the intraclass correlation coefficient ICC 1 , calculated as between-subject variance divided by the sum of between- and within-subject variance . Bias corrected and accelerated bootstrap confidence intervals were calculated for between- and within-subject CVs and ICC 1 using 10, bootstrap samples .
In order to explore time and temperature effects during simulation of transport in serum and plasma samples and the effect of tube type and number of thawing cycles on metabolite concentration in serum samples, nonparametric Friedman tests were performed . Significant changes in concentration during the Friedman tests were afterwards tested in a pairwise manner against the reference 0 h with Wilcoxon signed-rank tests.
Similarly, significant differences in metabolite concentrations between serum and plasma baseline samples serum W and plasma-direct were tested in a pairwise manner using Wilcoxon signed-rank tests. A power analysis was performed for each metabolite for selected settings. The minimum concentration change between two paired settings e.
To account for the reduced power of the non-parametric Wilcoxon signed-rank as compared to a parametric paired t-test, an asymptotic relative efficiency of 0. Results are given in table S3. Samples taken from 20 healthy human subjects on three different days were used to calculate the intraclass correlation coefficient ICC , the within-person CV WCV and the between-person CV BCV for metabolites in serum figure 1 and plasma figure S1.
In serum, the median WCV was 0. Median WCV values of metabolite subclasses were concordant in serum and plasma samples, whereas median BCV values were slightly higher in serum samples.
Introduction to metabolomics and its applications in ophthalmology
Both within- and between-person variance contribute to the ICC. Accordingly, the median ICC was higher in serum as compared to plasma samples. A Metabolites with median ICC-values below 0. The median ICC in plasma was 0. In serum samples, the median ICC was 0. Regarding metabolite subclasses, reliability was lowest for acylcarnitines both in serum and plasma, as hydroxyacylcarnitines showed low reliability due to low concentrations.
Only sarcosine showed significantly lower concentrations in serum compared to plasma samples table S5. There was no evidence of any effects of gender and last meal composition before each overnight fasting on metabolite concentrations data not shown. We simulated different ways of sample collection and shipment that might occur in multicenter clinical trials: Transport of non-centrifuged samples on cool packs CP within 3, 6 and 24 hours plasma and serum or at room temperature RT for 24 h plasma only.
Figure 3 shows examples of significantly altered metabolite concentrations in plasma. Stars in boxplots indicate significant difference in concentration compared to baseline 0 h. The compound class of amino acids was the least stable metabolite class. In the compound class of glycerophospholipids, all metabolites had stable concentrations on CP for at least 24 h. The compound class of sphingolipids was the most stable one, as all analyzed sphingolipid concentrations were stable on CP and at RT for at least 24 h.
In plasma, a larger number of metabolites had stable concentrations for 24 h on CP as compared to serum. In total, of serum metabolites had stable concentrations on CP for at least 24 h.
Figure 4 shows examples of altered metabolite concentrations in serum. In concordance with plasma samples, the group of amino acids was the least stable one in serum metabolites. Example of A - C increasing and D decreasing metabolite concentration during transportation simulation of serum samples on cool packs CP. In multicenter studies, the use of serum gel-barrier tubes is preferred due to the easy handling at the blood collection site and persistent separation of serum during shipment of tubes. In this study, possible interactions of the gel with the metabolites in human serum samples were analyzed.
Blood was drawn into serum gel-barrier tubes and serum W tubes with clotting activator only control. After centrifugation, whole gel-barrier tubes and separated serum from serum W tubes were frozen until analysis. Methionine sulfoxide showed significantly increased concentrations in serum gel-barrier tubes as compared to serum W see figure 5. For all other investigated metabolites, serum gel-barrier tubes had no significant effect on concentration, compared to serum W tubes. Stars in boxplots indicate significant differences in concentration between methionine sulfoxide in serum W tubes with clotting activator and serum gel-barrier tubes.
Repeated freezing and thawing of samples is sometimes inevitable due to a limited number of aliquots. When serum samples were subjected to two freeze-thaw cycles, most metabolite concentrations were stable. Except methionine sulfoxide, all investigated metabolites of this panel maintained stable concentrations after one freeze-thaw cycle. The minimal effect size i. This information should be informative for planning future studies. This study was not powered to significantly detect smaller effects than those displayed.
The analysis of three fasting blood samples within a 14 day period showed that most of the metabolite levels remained stable within the same individual. An ICC above 0. Compared to a population-based approach, our ICC values were artificially low due to the homogeneous study cohort. Metabolites with good reliability included saturated short- and medium acylcarnitines, amino acids, biogenic amines, glycerophospholipids, sphingolipids and hexose.
- Document Preview.
- IUTAM Symposium on Mesoscopic Dynamics of Fracture Process and Materials Strength: Proceedings of the IUTAM Symposium held in Osaka, Japan, 6–11 July 2003;
- Rationality and Dynamic Choice: Foundational Explorations!
- Navigation Bar.
- Mass Spectrometry in Grape and Wine Chemistry (Wiley - Interscience Series on Mass Spectrometry).
These findings are in line with Floegel et al. Creatinine was found to have the lowest WCV and highest reliability in plasma. In muscle cells, the ring closing of creatine forms creatinine in a chemical equilibrium . As the quantity of creatine per unit of skeletal muscle and the reaction to creatinine is constant  , the produced amount of creatinine is dependent on the muscle mass.
Concentration of creatinine is therefore constant and within-person variability is relatively low in healthy people, as long as the muscle mass remains constant. Reliability of metabolites was slightly higher in serum compared to plasma. This finding is not in line with Yu et al. Ladenson et al.
Publications » The Yost Research Group » University of Florida
Although glucose concentration was lower in plasma samples, the difference between serum and plasma levels of glucose was not significant in our study. We found no effect of sex and last meal composition before overnight fasting in contrast to other studies  — . This could be explained by the sample size and the unbalanced ratio of men to women. Our study goes beyond previous investigations by using a broader metabolite panel and measuring three samples per individual at the same time, in a well-established and validated targeted metabolomics approach.
The observed high reliability of amino acids was consistent with the results of Floegel et al. However, reliability for six additional amino acids was shown in our study. The good repeatability of phospholipid measurements in plasma samples observed by Ma et al. Classification of reliability of acylcarnitines, glycerophospholipids, sphingolipids and hexose was concordant with that of Floegel et al. Widjaja et al. No analytical duplicates were measured. Therefore, it was not possible to further divide the total within-person variability into biological and analytical variability.
However, regarding the between and within-plate CVs derived from repeated measurements of reference samples table S2 , analytical variance can be assumed to be low. Only samples of healthy subjects were analyzed, therefore variability in non-healthy patients could vary from these results. In sum, we found that most metabolite concentrations in a targeted approach are highly stable over time in a fasted individual.
The WCVs obtained in this study can help to judge whether single measurements of a metabolite are sufficient to describe differences between groups of human subjects. The majority of the analyzed metabolites had stable concentrations for at least 24 h on CP plasma and serum and at RT plasma.
Related Metabolome Analysis: An Introduction (Wiley - Interscience Series on Mass Spectrometry)
Copyright 2019 - All Right Reserved