Multivariate strategy for the sample selection and integration of multi-batch data in metabolomics

Introduction

Availability of large cohorts of samples with related metadata provides scientists with extensive material for studies. At the same time, recent development of modern high-throughput ‘omics’ technologies, including metabolomics, has resulted in the potential for analysis of large sample sizes. Representative subset selection becomes critical for selection of samples from bigger cohorts and their division into analytical batches. This especially holds true when relative quantification of compound levels is used.

Objectives

We present a multivariate strategy for representative sample selection and integration of results from multi-batch experiments in metabolomics.

Methods

Multivariate characterization was applied for design of experiment based sample selection and subsequent subdivision into four analytical batches which were analyzed on different days by metabolomics profiling using gas-chromatography time-of-flight mass spectrometry (GC–TOF–MS). For each batch OPLS-DA^® was used and its p(corr) vectors were averaged to obtain combined metabolic profile. Jackknifed standard errors were used to calculate confidence intervals for each metabolite in the average p(corr) profile.

Results

A combined, representative metabolic profile describing differences between systemic lupus erythematosus (SLE) patients and controls was obtained and used for elucidation of metabolic pathways that could be disturbed in SLE.

Conclusion

Design of experiment based representative sample selection ensured diversity and minimized bias that could be introduced at this step. Combined metabolic profile enabled unified analysis and interpretation.

     

KniMet: a pipeline for the processing of chromatography–mass spectrometry metabolomics data

Introduction

Data processing is one of the biggest problems in metabolomics, given the high number of samples analyzed and the need of multiple software packages for each step of the processing workflow.

Objectives

Merge in the same platform the steps required for metabolomics data processing.

Methods

KniMet is a workflow for the processing of mass spectrometry-metabolomics data based on the KNIME Analytics platform.

Results

The approach includes key steps to follow in metabolomics data processing: feature filtering, missing value imputation, normalization, batch correction and annotation.

Conclusion

KniMet provides the user with a local, modular and customizable workflow for the processing of both GC–MS and LC–MS open profiling data.

     

Integration of multiple types of genetic markers for neuroblastoma may contribute to improved prediction of the overall survival

Background

Modern experimental techniques deliver data sets containing profiles of tens of thousands of potential molecular and genetic markers that can be used to improve medical diagnostics. Previous studies performed with three different experimental methods for the same set of neuroblastoma patients create opportunity to examine whether augmenting gene expression profiles with information on copy number variation can lead to improved predictions of patients survival. We propose methodology based on comprehensive cross-validation protocol, that includes feature selection within cross-validation loop and classification using machine learning. We also test dependence of results on the feature selection process using four different feature selection methods.

Results

The models utilising features selected based on information entropy are slightly, but significantly, better than those using features obtained with t-test. The synergy between data on genetic variation and gene expression is possible, but not confirmed. A slight, but statistically significant, increase of the predictive power of machine learning models has been observed for models built on combined data sets. It was found while using both out of bag estimate and in cross-validation performed on a single set of variables. However, the improvement was smaller and non-significant when models were built within full cross-validation procedure that included feature selection within cross-validation loop. Good correlation between performance of the models in the internal and external cross-validation was observed, confirming the robustness of the proposed protocol and results.

Conclusions

We have developed a protocol for building predictive machine learning models. The protocol can provide robust estimates of the model performance on unseen data. It is particularly well-suited for small data sets. We have applied this protocol to develop prognostic models for neuroblastoma, using data on copy number variation and gene expression. We have shown that combining these two sources of information may increase the quality of the models. Nevertheless, the increase is small and larger samples are required to reduce noise and bias arising due to overfitting.

Reviewers

This article was reviewed by Lan Hu, Tim Beissbarth and Dimitar Vassilev.

     

integRATE: a desirability-based data integration framework for the prioritization of candidate genes across heterogeneous omics and its application to preterm birth

Background

The integration of high-quality, genome-wide analyses offers a robust approach to elucidating genetic factors involved in complex human diseases. Even though several methods exist to integrate heterogeneous omics data, most biologists still manually select candidate genes by examining the intersection of lists of candidates stemming from analyses of different types of omics data that have been generated by imposing hard (strict) thresholds on quantitative variables, such as P-values and fold changes, increasing the chance of missing potentially important candidates.

Methods

To better facilitate the unbiased integration of heterogeneous omics data collected from diverse platforms and samples, we propose a desirability function framework for identifying candidate genes with strong evidence across data types as targets for follow-up functional analysis. Our approach is targeted towards disease systems with sparse, heterogeneous omics data, so we tested it on one such pathology: spontaneous preterm birth (sPTB).

Results

We developed the software integRATE, which uses desirability functions to rank genes both within and across studies, identifying well-supported candidate genes according to the cumulative weight of biological evidence rather than based on imposition of hard thresholds of key variables. Integrating 10 sPTB omics studies identified both genes in pathways previously suspected to be involved in sPTB as well as novel genes never before linked to this syndrome. integRATE is available as an R package on GitHub (https://github.com/haleyeidem/integRATE).

Conclusions

Desirability-based data integration is a solution most applicable in biological research areas where omics data is especially heterogeneous and sparse, allowing for the prioritization of candidate genes that can be used to inform more targeted downstream functional analyses.

     

Derivative component analysis for mass spectral serum proteomic profiles

Background

As a promising way to transform medicine, mass spectrometry based proteomics technologies have seen a great progress in identifying disease biomarkers for clinical diagnosis and prognosis. However, there is a lack of effective feature selection methods that are able to capture essential data behaviors to achieve clinical level disease diagnosis. Moreover, it faces a challenge from data reproducibility, which means that no two independent studies have been found to produce same proteomic patterns. Such reproducibility issue causes the identified biomarker patterns to lose repeatability and prevents it from real clinical usage.

Methods

In this work, we propose a novel machine-learning algorithm: derivative component analysis (DCA) for high-dimensional mass spectral proteomic profiles. As an implicit feature selection algorithm, derivative component analysis examines input proteomics data in a multi-resolution approach by seeking its derivatives to capture latent data characteristics and conduct de-noising. We further demonstrate DCA's advantages in disease diagnosis by viewing input proteomics data as a profile biomarker via integrating it with support vector machines to tackle the reproducibility issue, besides comparing it with state-of-the-art peers.

Results

Our results show that high-dimensional proteomics data are actually linearly separable under proposed derivative component analysis (DCA). As a novel multi-resolution feature selection algorithm, DCA not only overcomes the weakness of the traditional methods in subtle data behavior discovery, but also suggests an effective resolution to overcoming proteomics data's reproducibility problem and provides new techniques and insights in translational bioinformatics and machine learning. The DCA-based profile biomarker diagnosis makes clinical level diagnostic performances reproducible across different proteomic data, which is more robust and systematic than the existing biomarker discovery based diagnosis.

Conclusions

Our findings demonstrate the feasibility and power of the proposed DCA-based profile biomarker diagnosis in achieving high sensitivity and conquering the data reproducibility issue in serum proteomics. Furthermore, our proposed derivative component analysis suggests the subtle data characteristics gleaning and de-noising are essential in separating true signals from red herrings for high-dimensional proteomic profiles, which can be more important than the conventional feature selection or dimension reduction. In particular, our profile biomarker diagnosis can be generalized to other omics data for derivative component analysis (DCA)'s nature of generic data analysis.

     

Prediction of heme binding residues from protein sequences with integrative sequence profiles

Background

The heme-protein interactions are essential for various biological processes such as electron transfer, catalysis, signal transduction and the control of gene expression. The knowledge of heme binding residues can provide crucial clues to understand these activities and aid in functional annotation, however, insufficient work has been done on the research of heme binding residues from protein sequence information.

Methods

We propose a sequence-based approach for accurate prediction of heme binding residues by a novel integrative sequence profile coupling position specific scoring matrices with heme specific physicochemical properties. In order to select the informative physicochemical properties, we design an intuitive feature selection scheme by combining a greedy strategy with correlation analysis.

Results

Our integrative sequence profile approach for prediction of heme binding residues outperforms the conventional methods using amino acid and evolutionary information on the 5-fold cross validation and the independent tests.

Conclusions

The novel feature of an integrative sequence profile achieves good performance using a reduced set of feature vector elements.

     

Incorporating methylation genome information improves prediction accuracy for drug treatment responses

Background

Association studies using a single type of omics data have been successful in identifying disease-associated genetic markers, but the underlying mechanisms are unaddressed. To provide a possible explanation of how these genetic factors affect the disease phenotype, integration of multiple omics data is needed.

Results

We propose a novel method, LIPID (likelihood inference proposal for indirect estimation), that uses both single nucleotide polymorphism (SNP) and DNA methylation data jointly to analyze the association between a trait and SNPs. The total effect of SNPs is decomposed into direct and indirect effects, where the indirect effects are the focus of our investigation. Simulation studies show that LIPID performs better in various scenarios than existing methods. Application to the GAW20 data also leads to encouraging results, as the genes identified appear to be biologically relevant to the phenotype studied.

Conclusions

The proposed LIPID method is shown to be meritorious in extensive simulations and in real-data analyses.

     

A lost opportunity for science: journals promote data sharing in metabolomics but do not enforce it

Introduction

Data sharing is being increasingly required by journals and has been heralded as a solution to the ‘replication crisis’.

Objectives

(i) Review data sharing policies of journals publishing the most metabolomics papers associated with open data and (ii) compare these journals’ policies to those that publish the most metabolomics papers.

Methods

A PubMed search was used to identify metabolomics papers. Metabolomics data repositories were manually searched for linked publications.

Results

Journals that support data sharing are not necessarily those with the most papers associated to open metabolomics data.

Conclusion

Further efforts are required to improve data sharing in metabolomics.

     

Predicting adverse drug reactions through interpretable deep learning framework

Background

Adverse drug reactions (ADRs) are unintended and harmful reactions caused by normal uses of drugs. Predicting and preventing ADRs in the early stage of the drug development pipeline can help to enhance drug safety and reduce financial costs.

Methods

In this paper, we developed machine learning models including a deep learning framework which can simultaneously predict ADRs and identify the molecular substructures associated with those ADRs without defining the substructures a-priori.

Results

We evaluated the performance of our model with ten different state-of-the-art fingerprint models and found that neural fingerprints from the deep learning model outperformed all other methods in predicting ADRs. Via feature analysis on drug structures, we identified important molecular substructures that are associated with specific ADRs and assessed their associations via statistical analysis.

Conclusions

The deep learning model with feature analysis, substructure identification, and statistical assessment provides a promising solution for identifying risky components within molecular structures and can potentially help to improve drug safety evaluation.

     

Untargeted metabolomics suffers from incomplete raw data processing

Introduction

Untargeted metabolomics is a powerful tool for biological discoveries. To analyze the complex raw data, significant advances in computational approaches have been made, yet it is not clear how exhaustive and reliable the data analysis results are.

Objectives

Assessment of the quality of raw data processing in untargeted metabolomics.

Methods

Five published untargeted metabolomics studies, were reanalyzed.

Results

Omissions of at least 50 relevant compounds from the original results as well as examples of representative mistakes were reported for each study.

Conclusion

Incomplete raw data processing shows unexplored potential of current and legacy data.

     

Multi-resolution independent component analysis for high-performance tumor classification and biomarker discovery

Background

Although high-throughput microarray based molecular diagnostic technologies show a great promise in cancer diagnosis, it is still far from a clinical application due to its low and instable sensitivities and specificities in cancer molecular pattern recognition. In fact, high-dimensional and heterogeneous tumor profiles challenge current machine learning methodologies for its small number of samples and large or even huge number of variables (genes). This naturally calls for the use of an effective feature selection in microarray data classification.

Methods

We propose a novel feature selection method: multi-resolution independent component analysis (MICA) for large-scale gene expression data. This method overcomes the weak points of the widely used transform-based feature selection methods such as principal component analysis (PCA), independent component analysis (ICA), and nonnegative matrix factorization (NMF) by avoiding their global feature-selection mechanism. In addition to demonstrating the effectiveness of the multi-resolution independent component analysis in meaningful biomarker discovery, we present a multi-resolution independent component analysis based support vector machines (MICA-SVM) and linear discriminant analysis (MICA-LDA) to attain high-performance classifications in low-dimensional spaces.

Results

We have demonstrated the superiority and stability of our algorithms by performing comprehensive experimental comparisons with nine state-of-the-art algorithms on six high-dimensional heterogeneous profiles under cross validations. Our classification algorithms, especially, MICA-SVM, not only accomplish clinical or near-clinical level sensitivities and specificities, but also show strong performance stability over its peers in classification. Software that implements the major algorithm and data sets on which this paper focuses are freely available at https://sites.google.com/site/heyaumapbc2011/.

Conclusions

This work suggests a new direction to accelerate microarray technologies into a clinical routine through building a high-performance classifier to attain clinical-level sensitivities and specificities by treating an input profile as a ‘profile-biomarker’. The multi-resolution data analysis based redundant global feature suppressing and effective local feature extraction also have a positive impact on large scale ‘omics’ data mining.

     

Incorporating methylation genome information improves prediction accuracy for drug treatment responses

Background

An accumulation of evidence has revealed the important role of epigenetic factors in explaining the etiopathogenesis of human diseases. Several empirical studies have successfully incorporated methylation data into models for disease prediction. However, it is still a challenge to integrate different types of omics data into prediction models, and the contribution of methylation information to prediction remains to be fully clarified.

Results

A stratified drug-response prediction model was built based on an artificial neural network to predict the change in the circulating triglyceride level after fenofibrate intervention. Associated single-nucleotide polymorphisms (SNPs), methylation of selected cytosine-phosphate-guanine (CpG) sites, age, sex, and smoking status, were included as predictors. The model with selected SNPs achieved a mean 5-fold cross-validation prediction error rate of 43.65%. After adding methylation information into the model, the error rate dropped to 41.92%. The combination of significant SNPs, CpG sites, age, sex, and smoking status, achieved the lowest prediction error rate of 41.54%.

Conclusions

Compared to using SNP data only, adding methylation data in prediction models slightly improved the error rate; further prediction error reduction is achieved by a combination of genome, methylation genome, and environmental factors.

     

How the chemical features of molecules may have addressed the settlement of metabolic steps

Introduction

While the evolutionary adaptation of enzymes to their own substrates is a well assessed and rationalized field, how molecules have been originally selected in order to initiate and assemble convenient metabolic pathways is a fascinating, but still debated argument.

Objectives

Aim of the present study is to give a rationale for the preferential selection of specific molecules to generate metabolic pathways.

Methods

The comparison of structural features of molecules, through an inductive methodological approach, offer a reading key to cautiously propose a determining factor for their metabolic recruitment.

Results

Starting with some commonplaces occurring in the structural representation of relevant carbohydrates, such as glucose, fructose and ribose, arguments are presented in associating stable structural determinants of these molecules and their peculiar occurrence in metabolic pathways.

Conclusions

Among other possible factors, the reliability of the structural asset of a molecule may be relevant or its selection among structurally and, a priori, functionally similar molecules.

     

The non-negative matrix factorization toolbox for biological data mining

Background

Non-negative matrix factorization (NMF) has been introduced as an important method for mining biological data. Though there currently exists packages implemented in R and other programming languages, they either provide only a few optimization algorithms or focus on a specific application field. There does not exist a complete NMF package for the bioinformatics community, and in order to perform various data mining tasks on biological data.

Results

We provide a convenient MATLAB toolbox containing both the implementations of various NMF techniques and a variety of NMF-based data mining approaches for analyzing biological data. Data mining approaches implemented within the toolbox include data clustering and bi-clustering, feature extraction and selection, sample classification, missing values imputation, data visualization, and statistical comparison.

Conclusions

A series of analysis such as molecular pattern discovery, biological process identification, dimension reduction, disease prediction, visualization, and statistical comparison can be performed using this toolbox.

     

Pseudo current density maps of electrophysiological heart,nerve or brain function and their physical basis

Background

In recent years the visualization of biomagnetic measurement data by so-called pseudo current density maps or Hosaka-Cohen (HC) transformations became popular.

Methods

The physical basis of these intuitive maps is clarified by means of analytically solvable problems.

Results

Examples in magnetocardiography, magnetoencephalography and magnetoneurography demonstrate the usefulness of this method.

Conclusion

Hardware realizations of the HC-transformation and some similar transformations are discussed which could advantageously support cross-platform comparability of biomagnetic measurements.

     

Classification of motor imagery tasks for BCI with multiresolution analysis and multiobjective feature selection

Background

Brain-computer interfacing (BCI) applications based on the classification of electroencephalographic (EEG) signals require solving high-dimensional pattern classification problems with such a relatively small number of training patterns that curse of dimensionality problems usually arise. Multiresolution analysis (MRA) has useful properties for signal analysis in both temporal and spectral analysis, and has been broadly used in the BCI field. However, MRA usually increases the dimensionality of the input data. Therefore, some approaches to feature selection or feature dimensionality reduction should be considered for improving the performance of the MRA based BCI.

Methods

This paper investigates feature selection in the MRA-based frameworks for BCI. Several wrapper approaches to evolutionary multiobjective feature selection are proposed with different structures of classifiers. They are evaluated by comparing with baseline methods using sparse representation of features or without feature selection.

Results and conclusion

The statistical analysis, by applying the Kolmogorov-Smirnoff and Kruskal–Wallis tests to the means of the Kappa values evaluated by using the test patterns in each approach, has demonstrated some advantages of the proposed approaches. In comparison with the baseline MRA approach used in previous studies, the proposed evolutionary multiobjective feature selection approaches provide similar or even better classification performances, with significant reduction in the number of features that need to be computed.

     

Identification of glioblastoma gene prognosis modules based on weighted gene co-expression network analysis

Background

Glioblastoma multiforme, the most prevalent and aggressive brain tumour, has a poor prognosis. The molecular mechanisms underlying gliomagenesis remain poorly understood. Therefore, molecular research, including various markers, is necessary to understand the occurrence and development of glioma.

Method

Weighted gene co-expression network analysis (WGCNA) was performed to construct a gene co-expression network in TCGA glioblastoma samples. Gene ontology (GO) and pathway-enrichment analysis were used to identify significance of gene modules. Cox proportional hazards regression model was used to predict outcome of glioblastoma patients.

Results

We performed weighted gene co-expression network analysis (WGCNA) and identified a gene module (yellow module) related to the survival time of TCGA glioblastoma samples. Then, 228 hub genes were calculated based on gene significance (GS) and module significance (MS). Four genes (OSMR + SOX21?+?MED10?+?PTPRN) were selected to construct a Cox proportional hazards regression model with high accuracy (AUC?=?0.905). The prognostic value of the Cox proportional hazards regression model was also confirmed in GSE16011 dataset (GBM: n?=?156).

Conclusion

We developed a promising mRNA signature for estimating overall survival in glioblastoma patients.