Separating the wheat from the chaff: a prioritisation pipeline for the analysis of metabolomics datasets

Liquid Chromatography Mass Spectrometry (LC-MS) is a powerful and widely applied method for the study of biological systems, biomarker discovery and pharmacological interventions. LC-MS measurements are, however, significantly complicated by several technical challenges, including: (1) ionisation suppression/enhancement, disturbing the correct quantification of analytes, and (2) the detection of large amounts of separate derivative ions, increasing the complexity of the spectra, but not their information content. Here we introduce an experimental and analytical strategy that leads to robust metabolome profiles in the face of these challenges. Our method is based on rigorous filtering of the measured signals based on a series of sample dilutions. Such data sets have the additional characteristic that they allow a more robust assessment of detection signal quality for each metabolite. Using our method, almost 80% of the recorded signals can be discarded as uninformative, while important information is retained. As a consequence, we obtain a broader understanding of the information content of our analyses and a better assessment of the metabolites detected in the analyzed data sets. We illustrate the applicability of this method using standard mixtures, as well as cell extracts from bacterial samples. It is evident that this method can be applied in many types of LC-MS analyses and more specifically in untargeted metabolomics.     

A geometric approach for the alignment of liquid chromatography-mass spectrometry data

MOTIVATION: Liquid chromatography coupled to mass spectrometry (LC-MS) and combined with tandem mass spectrometry (LC-MS/MS) have become a prominent tool for the analysis of complex proteomic samples. An important step in a typical workflow is the combination of results from multiple LC-MS experiments to improve confidence in the obtained measurements or to compare results from different samples. To do so, a suitable mapping or alignment between the data sets needs to be estimated. The alignment has to correct for variations in mass and elution time which are present in all mass spectrometry experiments. RESULTS: We propose a novel algorithm to align LC-MS samples and to match corresponding ion species across samples. Our algorithm matches landmark signals between two data sets using a geometric technique based on pose clustering. Variations in mass and retention time are corrected by an affine dewarping function estimated from matched landmarks. We use the pairwise dewarping in an algorithm for aligning multiple samples. We show that our pose clustering approach is fast and reliable as compared to previous approaches. It is robust in the presence of noise and able to accurately align samples with only few common ion species. In addition, we can easily handle different kinds of LC-MS data and adopt our algorithm to new mass spectrometry technologies. AVAILABILITY: This algorithm is implemented as part of the OpenMS software library for shotgun proteomics and available under the Lesser GNU Public License (LGPL) at www.openms.de.     

Life-LCA: assessing the environmental impacts of a human being—challenges and perspectives

Purpose

The life-cycle assessment (LCA) method is typically applied to products, but the potential and demand for extending its use also to other applications are high. In this respect, this paper proposes an LCA concept to be used for the assessment of human beings as new study objects, namely Life-LCA. Key challenges of such a new approach and potential solutions for those are identified and discussed.

Methods

The Life-LCA concept was developed based on a detailed desktop research. Several Life-LCA-specific challenges were identified and categorized under three research questions. One of these questions focusses on the conceptual design of a Life-LCA method while the others are addressing operational issues, which are the definition of the new study system and the practical assessment of complex human consumption behaviors. Methodological solutions are proposed, e.g., based on suggestions provided in the existing methods product LCA and organizational LCA (O-LCA).

Results and discussion

Conceptual challenges arise from the general diversity, complexity, and temporal development of human lives and consumption behaviors. We introduce Life-LCA as a two-dimensional method that covers both, the new human life cycle (dimension 1) and the life cycle of the consumed products (dimension 2). Furthermore, the two types Individual Life-LCA and Lifestyle-LCA are differentiated. Especially, the definition of a general system boundary for Life-LCA and data collection and evaluation face many operational challenges. For example, the social behavior of human beings is a new factor to be considered which causes new allocation problems in LCA. Moreover, the high demand for aggregated LCA data requires specific rules for data collection and evaluation as well as a new bottom-up product clustering scheme.

Conclusions

Life-LCA, either used for the assessment of individual lives or lifestyles, has the potential to raise environmental awareness of people by making their specific environmental impacts comprehensively measurable and thus, tangible. However, many challenges need to be solved in future interdisciplinary research to develop a robust and applicable method. This paper conceptualizes such an approach and proposes solutions that can serve as a framework for ongoing method development.

     

Quality control metrics for LC-MS feature detection tools demonstrated on Saccharomyces cerevisiae proteomic profiles

Quantitative proteomic profiling using liquid chromatography-mass spectrometry is emerging as an important tool for biomarker discovery, prompting development of algorithms for high-throughput peptide feature detection in complex samples. However, neither annotated standard data sets nor quality control metrics currently exist for assessing the validity of feature detection algorithms. We propose a quality control metric, Mass Deviance, for assessing the accuracy of feature detection tools. Because the Mass Deviance metric is derived from the natural distribution of peptide masses, it is machine- and proteome-independent and enables assessment of feature detection tools in the absence of completely annotated data sets. We validate the use of Mass Deviance with a second, independent metric that is based on isotopic distributions, demonstrating that we can use Mass Deviance to identify aberrant features with high accuracy. We then demonstrate the use of independent metrics in tandem as a robust way to evaluate the performance of peptide feature detection algorithms. This work is done on complex LC-MS profiles of Saccharomyces cerevisiae which present a significant challenge to peptide feature detection algorithms.     

A Hybrid Approach of Gene Sets and Single Genes for the Prediction of Survival Risks with Gene Expression Data

Accumulated biological knowledge is often encoded as gene sets, collections of genes associated with similar biological functions or pathways. The use of gene sets in the analyses of high-throughput gene expression data has been intensively studied and applied in clinical research. However, the main interest remains in finding modules of biological knowledge, or corresponding gene sets, significantly associated with disease conditions. Risk prediction from censored survival times using gene sets hasn’t been well studied. In this work, we propose a hybrid method that uses both single gene and gene set information together to predict patient survival risks from gene expression profiles. In the proposed method, gene sets provide context-level information that is poorly reflected by single genes. Complementarily, single genes help to supplement incomplete information of gene sets due to our imperfect biomedical knowledge. Through the tests over multiple data sets of cancer and trauma injury, the proposed method showed robust and improved performance compared with the conventional approaches with only single genes or gene sets solely. Additionally, we examined the prediction result in the trauma injury data, and showed that the modules of biological knowledge used in the prediction by the proposed method were highly interpretable in biology. A wide range of survival prediction problems in clinical genomics is expected to benefit from the use of biological knowledge.     

Improving detection of copy-number variation by simultaneous bias correction and read-depth segmentation

Structural variation is an important class of genetic variation in mammals. High-throughput sequencing (HTS) technologies promise to revolutionize copy-number variation (CNV) detection but present substantial analytic challenges. Converging evidence suggests that multiple types of CNV-informative data (e.g. read-depth, read-pair, split-read) need be considered, and that sophisticated methods are needed for more accurate CNV detection. We observed that various sources of experimental biases in HTS confound read-depth estimation, and note that bias correction has not been adequately addressed by existing methods. We present a novel read-depth–based method, GENSENG, which uses a hidden Markov model and negative binomial regression framework to identify regions of discrete copy-number changes while simultaneously accounting for the effects of multiple confounders. Based on extensive calibration using multiple HTS data sets, we conclude that our method outperforms existing read-depth–based CNV detection algorithms. The concept of simultaneous bias correction and CNV detection can serve as a basis for combining read-depth with other types of information such as read-pair or split-read in a single analysis. A user-friendly and computationally efficient implementation of our method is freely available.     

Analysis of single-molecule mechanical recordings: application to acto-myosin interactions

Several laboratories have now developed methods to make single-molecule mechanical recordings from interacting pairs of biological molecules. The mechanical work done (product of force and distance) by a single biomolecular interaction is usually of the same order as thermal energy. Recordings made from non-processive, intermittently interacting, molecular motors such as acto-myosin therefore contain a large background of thermal noise. We have applied Page's test to analyse mechanical interactions between muscle myosin II's and F-actin recorded using an optical tweezers based single-molecule mechanical transducer. We compare Page's test with other variance-based methods and find it to be a robust method for analysing both simulated and real data sets. We discuss some of the problems associated with automatic detection of transient mechanical events in noisy data signals, and show that if the start and end points of individual events are known accurately then the events may be synchronised and combined to give more detailed information about different mechanical states.     

Mining phosphopeptide signals in liquid chromatography-mass spectrometry data for protein phosphorylation analysis

Protein phosphorylation is a key post-translational modification that governs biological processes. Despite the fact that a number of analytical strategies have been exploited for the characterization of protein phosphorylation, the identification of protein phosphorylation sites is still challenging. We proposed here an alternative approach to mine phosphopeptide signals generated from a mixture of proteins when liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis is involved. The approach combined dephosphorylation reaction, accurate mass measurements from a quadrupole/time-of-flight mass spectrometer, and a computing algorithm to differentiate possible phosphopeptide signals obtained from the LC-MS analyses by taking advantage of the mass shift generated by alkaline phosphatase treatment. The retention times and m/z values of these selected LC-MS signals were used to facilitate subsequent LC-MS/MS experiments for phosphorylation site determination. Unlike commonly used neutral loss scan experiments for phosphopeptide detection, this strategy may not bias against tyrosine-phosphorylated peptides. We have demonstrated the applicability of this strategy to sequence more, in comparison with conventional data-dependent LC-MS/MS experiments, phosphopeptides in a mixture of alpha- and beta-caseins. The analytical scheme was applied to characterize the nasopharyngeal carcinoma (NPC) cellular phosphoproteome and yielded 221 distinct phosphorylation sites. Our data presented in this paper demonstrated the merits of computation in mining phosphopeptide signals from a complex mass spectrometric data set.     

Automatically detecting the wild giant panda using deep learning with context and species distribution model

The giant panda is a flagship species in ecological conservation. The infrared camera trap is an effective tool for monitoring the giant panda. Images captured by infrared camera traps must be accurately recognized before further statistical analyses can be implemented. Previous research has demonstrated that spatiotemporal and positional contextual information and the species distribution model (SDM) can improve image detection accuracy, especially for difficult-to-see images. Difficult-to-see images include those in which individual animals are only partially observed and it is challenging for the model to detect those individuals. By utilizing the attention mechanism, we developed a unique method based on deep learning that incorporates object detection, contextual information, and the SDM to achieve better detection performance in difficult-to-see images. We obtained 1169 images of the wild giant panda and divided them into a training set and a test set in a 4:1 ratio. Model assessment metrics showed that our proposed model achieved an overall performance of 98.1% in mAP_0.5 and 82.9% in recall on difficult-to-see images. Our research demonstrated that the fine-grained multimodal-fusing method applied to monitoring giant pandas in the wild can better detect the difficult-to-see panda images to enhance the wildlife monitoring system.     

Double-labeled donor probe can enhance the signal of fluorescence resonance energy transfer (FRET) in detection of nucleic acid hybridization

A set of fluorescently-labeled DNA probes that hybridize with the target RNA and produce fluorescence resonance energy transfer (FRET) signals can be utilized for the detection of specific RNA. We have developed probe sets to detect and discriminate single-strand RNA molecules of plant viral genome, and sought a method to improve the FRET signals to handle in vivo applications. Consequently, we found that a double-labeled donor probe labeled with Bodipy dye yielded a remarkable increase in fluorescence intensity compared to a single-labeled donor probe used in an ordinary FRET. This double-labeled donor system can be easily applied to improve various FRET probes since the dependence upon sequence and label position in enhancement is not as strict. Furthermore this method could be applied to other nucleic acid substances, such as oligo RNA and phosphorothioate oligonucleotides (S-oligos) to enhance FRET signal. Although the double-labeled donor probes labeled with a variety of fluorophores had unexpected properties (strange UV-visible absorption spectra, decrease of intensity and decay of donor fluorescence) compared with single-labeled ones, they had no relation to FRET enhancement. This signal amplification mechanism cannot be explained simply based on our current results and knowledge of FRET. Yet it is possible to utilize this double-labeled donor system in various applications of FRET as a simple signal-enhancement method.     

MSClust: a tool for unsupervised mass spectra extraction of chromatography-mass spectrometry ion-wise aligned data

Mass peak alignment (ion-wise alignment) has recently become a popular method for unsupervised data analysis in untargeted metabolic profiling. Here we present MSClust-a software tool for analysis GC-MS and LC-MS datasets derived from untargeted profiling. MSClust performs data reduction using unsupervised clustering and extraction of putative metabolite mass spectra from ion-wise chromatographic alignment data. The algorithm is based on the subtractive fuzzy clustering method that allows unsupervised determination of a number of metabolites in a data set and can deal with uncertain memberships of mass peaks in overlapping mass spectra. This approach is based purely on the actual information present in the data and does not require any prior metabolite knowledge. MSClust can be applied for both GC-MS and LC-MS alignment data sets. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1007/s11306-011-0368-2) contains supplementary material, which is available to authorized users.     

Smoothness without Smoothing: Why Gaussian Naive Bayes Is Not Naive for Multi-Subject Searchlight Studies

Spatial smoothness is helpful when averaging fMRI signals across multiple subjects, as it allows different subjects'' corresponding brain areas to be pooled together even if they are slightly misaligned. However, smoothing is usually not applied when performing multivoxel pattern-based analyses (MVPA), as it runs the risk of blurring away the information that fine-grained spatial patterns contain. It would therefore be desirable, if possible, to carry out pattern-based analyses which take unsmoothed data as their input but which produce smooth images as output. We show here that the Gaussian Naive Bayes (GNB) classifier does precisely this, when it is used in “searchlight” pattern-based analyses. We explain why this occurs, and illustrate the effect in real fMRI data. Moreover, we show that analyses using GNBs produce results at the multi-subject level which are statistically robust, neurally plausible, and which replicate across two independent data sets. By contrast, SVM classifiers applied to the same data do not generate a replication, even if the SVM-derived searchlight maps have smoothing applied to them. An additional advantage of GNB classifiers for searchlight analyses is that they are orders of magnitude faster to compute than more complex alternatives such as SVMs. Collectively, these results suggest that Gaussian Naive Bayes classifiers may be a highly non-naive choice for multi-subject pattern-based fMRI studies.     

Accounting for linkage disequilibrium in genome scans for selection without individual genotypes: The local score approach

Detecting genomic footprints of selection is an important step in the understanding of evolution. Accounting for linkage disequilibrium in genome scans increases detection power, but haplotype‐based methods require individual genotypes and are not applicable on pool‐sequenced samples. We propose to take advantage of the local score approach to account for linkage disequilibrium in genome scans for selection, cumulating (possibly small) signals from single markers over a genomic segment, to clearly pinpoint a selection signal. Using computer simulations, we demonstrate that this approach detects selection with higher power than several state‐of‐the‐art single‐marker, windowing or haplotype‐based approaches. We illustrate this on two benchmark data sets including individual genotypes, for which we obtain similar results with the local score and one haplotype‐based approach. Finally, we apply the local score approach to Pool‐Seq data obtained from a divergent selection experiment on behaviour in quail and obtain precise and biologically coherent selection signals: while competing methods fail to highlight any clear selection signature, our method detects several regions involving genes known to act on social responsiveness or autistic traits. Although we focus here on the detection of positive selection from multiple population data, the local score approach is general and can be applied to other genome scans for selection or other genomewide analyses such as GWAS.     

Disaster Risk Assessment Based on Variable Fuzzy Sets and Improved Information Diffusion Method

The predictive analysis of natural disasters and their consequences is challenging because of uncertainties and incomplete data. The present article studies the use of variable fuzzy sets (VFS) and the improved information diffusion method (IIDM) to construct a composite method. The proposed method aims to integrate multiple factors and quantification of uncertainties within a consistent system for catastrophic risk assessment. VFS have been selected because of several advantages. These sets can effectively eliminate the border effect on assessment result and monitor estimation standard error. The IIDM theory can obtain as much useful information as possible from the sample and improves system recognition accuracy. Both these methods contribute to a sound risk probability prediction of natural disasters. Furthermore, the application has been verified through flood and drought risk analyses in China, where the risks of different flood and drought grades have been obtained. The results indicate that the proposed method achieves a better detection stability of disaster risk with higher precision compared with other methods. The method is effective, practical, and applicable in disaster rescue and relief.     

Association analyses of repeated measures on triglyceride and high-density lipoprotein levels: insights from GAW20

Background

The GAW20 group formed on the theme of methods for association analyses of repeated measures comprised 4sets of investigators. The provided “real” data set included genotypes obtained from a human whole-genome association study based on longitudinal measurements of triglycerides (TGs) and high-density lipoprotein in addition to methylation levels before and after administration of fenofibrate. The simulated data set contained 200 replications of methylation levels and posttreatment TGs, mimicking the real data set.

Results

The different investigators in the group focused on the statistical challenges unique to family-based association analyses of phenotypes measured longitudinally and applied a wide spectrum of statistical methods such as linear mixed models, generalized estimating equations, and quasi-likelihood–based regression models. This article discusses the varying strategies explored by the group’s investigators with the common goal of improving the power to detect association with repeated measures of a phenotype.

Conclusions

Although it is difficult to identify a common message emanating from the different contributions because of the diversity in the issues addressed, the unifying theme of the contributions lie in the search for novel analytic strategies to circumvent the limitations of existing methodologies to detect genetic association.

     

An integrated, directed mass spectrometric approach for in-depth characterization of complex peptide mixtures

LC-MS/MS has emerged as the method of choice for the identification and quantification of protein sample mixtures. For very complex samples such as complete proteomes, the most commonly used LC-MS/MS method, data-dependent acquisition (DDA) precursor selection, is of limited utility. The limited scan speed of current mass spectrometers along with the highly redundant selection of the most intense precursor ions generates a bias in the pool of identified proteins toward those of higher abundance. A directed LC-MS/MS approach that alleviates the limitations of DDA precursor ion selection by decoupling peak detection and sequencing of selected precursor ions is presented. In the first stage of the strategy, all detectable peptide ion signals are extracted from high resolution LC-MS feature maps or aligned sets of feature maps. The selected features or a subset thereof are subsequently sequenced in sequential, non-redundant directed LC-MS/MS experiments, and the MS/MS data are mapped back to the original LC-MS feature map in a fully automated manner. The strategy, implemented on an LTQ-FT MS platform, allowed the specific sequencing of 2,000 features per analysis and enabled the identification of more than 1,600 phosphorylation sites using a single reversed phase separation dimension without the need for time-consuming prefractionation steps. Compared with conventional DDA LC-MS/MS experiments, a substantially higher number of peptides could be identified from a sample, and this increase was more pronounced for low intensity precursor ions.     

Gaussian process regression model for normalization of LC-MS data using scan-level information

Background

Differences in sample collection, biomolecule extraction, and instrument variability introduce bias to data generated by liquid chromatography coupled with mass spectrometry (LC-MS). Normalization is used to address these issues. In this paper, we introduce a new normalization method using the Gaussian process regression model (GPRM) that utilizes information from individual scans within an extracted ion chromatogram (EIC) of a peak. The proposed method is particularly applicable for normalization based on analysis order of LC-MS runs. Our method uses measurement variabilities estimated through LC-MS data acquired from quality control samples to correct for bias caused by instrument drift. Maximum likelihood approach is used to find the optimal parameters for the fitted GPRM. We review several normalization methods and compare their performance with GPRM.

Results

To evaluate the performance of different normalization methods, we consider LC-MS data from a study where metabolomic approach is utilized to discover biomarkers for liver cancer. The LC-MS data were acquired by analysis of sera from liver cancer patients and cirrhotic controls. In addition, LC-MS runs from a quality control (QC) sample are included to assess the run to run variability and to evaluate the ability of various normalization method in reducing this undesired variability. Also, ANOVA models are applied to the normalized LC-MS data to identify ions with intensity measurements that are significantly different between cases and controls.

Conclusions

One of the challenges in using label-free LC-MS for quantitation of biomolecules is systematic bias in measurements. Several normalization methods have been introduced to overcome this issue, but there is no universally applicable approach at the present time. Each data set should be carefully examined to determine the most appropriate normalization method. We review here several existing methods and introduce the GPRM for normalization of LC-MS data. Through our in-house data set, we show that the GPRM outperforms other normalization methods considered here, in terms of decreasing the variability of ion intensities among quality control runs.

     

Fine-scale landscape genetics of the American badger (Taxidea taxus): disentangling landscape effects and sampling artifacts in a poorly understood species

Landscape genetics is a powerful tool for conservation because it identifies landscape features that are important for maintaining genetic connectivity between populations within heterogeneous landscapes. However, using landscape genetics in poorly understood species presents a number of challenges, namely, limited life history information for the focal population and spatially biased sampling. Both obstacles can reduce power in statistics, particularly in individual-based studies. In this study, we genotyped 233 American badgers in Wisconsin at 12 microsatellite loci to identify alternative statistical approaches that can be applied to poorly understood species in an individual-based framework. Badgers are protected in Wisconsin owing to an overall lack in life history information, so our study utilized partial redundancy analysis (RDA) and spatially lagged regressions to quantify how three landscape factors (Wisconsin River, Ecoregions and land cover) impacted gene flow. We also performed simulations to quantify errors created by spatially biased sampling. Statistical analyses first found that geographic distance was an important influence on gene flow, mainly driven by fine-scale positive spatial autocorrelations. After controlling for geographic distance, both RDA and regressions found that Wisconsin River and Agriculture were correlated with genetic differentiation. However, only Agriculture had an acceptable type I error rate (3–5%) to be considered biologically relevant. Collectively, this study highlights the benefits of combining robust statistics and error assessment via simulations and provides a method for hypothesis testing in individual-based landscape genetics.     

Localization of water-soluble carbohydrates in wheat stems using imaging matrix-assisted laser desorption ionization mass spectrometry

The pool of endogenous water-soluble oligosaccharides found in the stems of wheat (Triticum aestivum) is being investigated as a potential indicator of grain yield. Techniques such as liquid chromatography with mass spectrometry (LC-MS) can profile these analytes but provide no spatial information regarding their distribution in the wheat stem. The imaging matrix-assisted laser desorption ionization (MALDI) mass spectrometry technique has not been utilized for the analysis of oligosaccharides in plant systems previously. Imaging MALDI mass spectrometry was used to analyse cross and longitudinal sections from the stems of Triticum aestivum. A range of oligosaccharides up to Hex(11) were observed. Water-soluble oligosaccharides were ionized as potassiated molecules, and found to be located in the stem pith that is retained predominantly around the inner stem wall. Imaging MALDI analyses provided spatial information on endogenous oligosaccharides present in wheat stems. The technique was found to offer comparable sensitivities for oligosaccharide detection to those of our established LC-MS method, and has potential for broad application in studying the in situ localization of other compound types in plant material.