A critical assessment of feature selection methods for biomarker discovery in clinical proteomics.

In this paper, we compare the performance of six different feature selection methods for LC-MS-based proteomics and metabolomics biomarker discovery-t test, the Mann-Whitney-Wilcoxon test (mww test), nearest shrunken centroid (NSC), linear support vector machine-recursive features elimination (SVM-RFE), principal component discriminant analysis (PCDA), and partial least squares discriminant analysis (PLSDA)-using human urine and porcine cerebrospinal fluid samples that were spiked with a range of peptides at different concentration levels. The ideal feature selection method should select the complete list of discriminating features that are related to the spiked peptides without selecting unrelated features. Whereas many studies have to rely on classification error to judge the reliability of the selected biomarker candidates, we assessed the accuracy of selection directly from the list of spiked peptides. The feature selection methods were applied to data sets with different sample sizes and extents of sample class separation determined by the concentration level of spiked compounds. For each feature selection method and data set, the performance for selecting a set of features related to spiked compounds was assessed using the harmonic mean of the recall and the precision (f-score) and the geometric mean of the recall and the true negative rate (g-score). We conclude that the univariate t test and the mww test with multiple testing corrections are not applicable to data sets with small sample sizes (n = 6), but their performance improves markedly with increasing sample size up to a point (n > 12) at which they outperform the other methods. PCDA and PLSDA select small feature sets with high precision but miss many true positive features related to the spiked peptides. NSC strikes a reasonable compromise between recall and precision for all data sets independent of spiking level and number of samples. Linear SVM-RFE performs poorly for selecting features related to the spiked compounds, even though the classification error is relatively low.

Mutations in the phospholipid remodeling gene SERAC1 impair mitochondrial function and intracellular cholesterol trafficking and cause dystonia and deafness.

Using exome sequencing, we identify SERAC1 mutations as the cause of MEGDEL syndrome, a recessive disorder of dystonia and deafness with Leigh-like syndrome, impaired oxidative phosphorylation and 3-methylglutaconic aciduria. We localized SERAC1 at the interface between the mitochondria and the endoplasmic reticulum in the mitochondria-associated membrane fraction that is essential for phospholipid exchange. A phospholipid analysis in patient fibroblasts showed elevated concentrations of phosphatidylglycerol-34:1 (where the species nomenclature denotes the number of carbon atoms in the two acyl chains:number of double bonds in the two acyl groups) and decreased concentrations of phosphatidylglycerol-36:1 species, resulting in an altered cardiolipin subspecies composition. We also detected low concentrations of bis(monoacyl-glycerol)-phosphate, leading to the accumulation of free cholesterol, as shown by abnormal filipin staining. Complementation of patient fibroblasts with wild-type human SERAC1 by lentiviral infection led to a decrease and partial normalization of the mean ratio of phosphatidylglycerol-34:1 to phosphatidylglycerol-36:1. Our data identify SERAC1 as a key player in the phosphatidylglycerol remodeling that is essential for both mitochondrial function and intracellular cholesterol trafficking.

The impact of delayed storage on the measured proteome and metabolome of human cerebrospinal fluid.

BACKGROUND: Because cerebrospinal fluid (CSF) is in close contact with diseased areas in neurological disorders, it is an important source of material in the search for molecular biomarkers. However, sample handling for CSF collected from patients in a clinical setting might not always be adequate for use in proteomics and metabolomics studies. METHODS: We left CSF for 0, 30, and 120 min at room temperature immediately after sample collection and centrifugation/removal of cells. At 2 laboratories CSF proteomes were subjected to tryptic digestion and analyzed by use of nano-liquid chromatography (LC) Orbitrap mass spectrometry (MS) and chipLC quadrupole TOF-MS. Metabolome analysis was performed at 3 laboratories by NMR, GC-MS, and LC-MS. Targeted analyses of cystatin C and albumin were performed by LC-tandem MS in the selected reaction monitoring mode. RESULTS: We did not find significant changes in the measured proteome and metabolome of CSF stored at room temperature after centrifugation, except for 2 peptides and 1 metabolite, 2,3,4-trihydroxybutanoic (threonic) acid, of 5780 identified peptides and 93 identified metabolites. A sensitive protein stability marker, cystatin C, was not affected. CONCLUSIONS: The measured proteome and metabolome of centrifuged human CSF is stable at room temperature for up to 2 hours. We cannot exclude, however, that changes undetectable with our current methodology, such as denaturation or proteolysis, might occur because of sample handling conditions. The stability we observed gives laboratory personnel at the collection site sufficient time to aliquot samples before freezing and storage at -80 °C.

The effect of preanalytical factors on stability of the proteome and selected metabolites in cerebrospinal fluid (CSF).

To standardize the use of cerebrospinal fluid (CSF) for biomarker research, a set of stability studies have been performed on porcine samples to investigate the influence of common sample handling procedures on proteins, peptides, metabolites and free amino acids. This study focuses at the effect on proteins and peptides, analyzed by applying label-free quantitation using microfluidics nanoscale liquid chromatography coupled with quadrupole time-of-flight mass spectrometry (chipLC-MS) as well as matrix-assisted laser desorption ionization Fourier transform ion cyclotron resonance mass spectrometry (MALDI-FT-ICR-MS) and Orbitrap LC-MS/MS to trypsin-digested CSF samples. The factors assessed were a 30 or 120 min time delay at room temperature before storage at -80 degrees C after the collection of CSF in order to mimic potential delays in the clinic (delayed storage), storage at 4 degrees C after trypsin digestion to mimic the time that samples remain in the cooled autosampler of the analyzer, and repeated freeze-thaw cycles to mimic storage and handling procedures in the laboratory. The delayed storage factor was also analyzed by gas chromatography mass spectrometry (GC-MS) and liquid chromatography mass spectrometry (LC-MS) for changes of metabolites and free amino acids, respectively. Our results show that repeated freeze/thawing introduced changes in transthyretin peptide levels. The trypsin digested samples left at 4 degrees C in the autosampler showed a time-dependent decrease of peak areas for peptides from prostaglandin D-synthase and serotransferrin. Delayed storage of CSF led to changes in prostaglandin D-synthase derived peptides as well as to increased levels of certain amino acids and metabolites. The changes of metabolites, amino acids and proteins in the delayed storage study appear to be related to remaining white blood cells. Our recommendations are to centrifuge CSF samples immediately after collection to remove white blood cells, aliquot, and then snap-freeze the supernatant in liquid nitrogen for storage at -80 degrees C. Preferably samples should not be left in the autosampler for more than 24 h and freeze/thaw cycles should be avoided if at all possible.

Optimized time alignment algorithm for LC-MS data: correlation optimized warping using component detection algorithm-selected mass chromatograms.

Correlation optimized warping (COW) based on the total ion current (TIC) is a widely used time alignment algorithm (COW-TIC). This approach works successfully on chromatograms containing few compounds and having a well-defined TIC. In this paper, we have combined COW with a component detection algorithm (CODA) to align LC-MS chromatograms containing thousands of biological compounds with overlapping chromatographic peaks, a situation where COW-TIC often fails. CODA is a variable selection procedure that selects mass chromatograms with low noise and low background (so-called "high-quality" mass chromatograms). High-quality mass chromatograms selected in each COW segment ensure that the same compounds (based on their mass and their retention time) are used in the two-dimensional benefit function of COW to obtain correct and optimal alignments (COW-CODA). The performance of the COW-CODA algorithm was evaluated on three types of complex data sets obtained from the LC-MS analysis of samples commonly used for biomarker discovery and compared to COW-TIC using a new global comparison method based on overlapping peak area: trypsin-digested serum obtained from cervical cancer patients, trypsin-digested serum from a single patient that was treated with varying preanalytical parameters (factorial design study), and urine from pregnant and nonpregnant women. While COW-CODA did result in minor misalignments in rare cases, it was clearly superior to the COW-TIC algorithm, especially when applied to highly variable chromatograms (factorial design, urine). The presented algorithm thus enables automatic time alignment and accurate peak matching of multiple LC-MS data sets obtained from complex body fluids that are often used for biomarker discovery.