Genomic computing. Explanatory analysis of plant expression profiling data using machine learning

Raiders of the Lost Ark — A Review About the Roots and Application of Artificial Intelligence

Artificial Intelligence (AI) receives prominent attention within the innovation context. It is the most promising technological invention in information technology. Nevertheless, Innovation and Technology Management (ITM) so far could not structure the AI field, which offers a disruptive innovative potential. Thus, this paper reviews and analyzes the ITM literature and explains the underlying structure of AI. The findings present two main streams of AI literature and, furthermore, explain how to categorize AI use cases. With our results, we assist ITM in explaining and adopting AI to business, which is a huge challenge for companies.

Ensemble machine learning modeling for the prediction of artemisinin resistance in malaria

Resistance in malaria is a growing concern affecting many areas of Sub-Saharan Africa and Southeast Asia. Since the emergence of artemisinin resistance in the late 2000s in Cambodia, research into the underlying mechanisms has been underway. The 2019 Malaria Challenge posited the task of developing computational models that address important problems in advancing the fight against malaria. The first goal was to accurately predict artemisinin drug resistance levels of Plasmodium falciparum isolates, as quantified by the IC 50. The second goal was to predict the parasite clearance rate of malaria parasite isolates based on in vitro transcriptional profiles. In this work, we develop machine learning models using novel methods for transforming isolate data and handling the tens of thousands of variables that result from these data transformation exercises. This is demonstrated by using massively parallel processing of the data vectorization for use in scalable machine learning. In addition, we show the utility of ensemble machine learning modeling for highly effective predictions of both goals of this challenge. This is demonstrated by the use of multiple machine learning algorithms combined with various scaling and normalization preprocessing steps. Then, using a voting ensemble, multiple models are combined to generate a final model prediction.

Ensemble machine learning modeling for the prediction of artemisinin resistance in malaria

Resistance in malaria is a growing concern affecting many areas of Sub-Saharan Africa and Southeast Asia. Since the emergence of artemisinin resistance in the late 2000s in Cambodia, research into the underlying mechanisms has been underway. The 2019 Malaria Challenge posited the task of developing computational models that address important problems in advancing the fight against malaria. The first goal was to accurately predict artemisinin drug resistance levels of Plasmodium falciparum isolates, as quantified by the IC 50. The second goal was to predict the parasite clearance rate of malaria parasite isolates based on in vitro transcriptional profiles. In this work, we develop machine learning models using novel methods for transforming isolate data and handling the tens of thousands of variables that result from these data transformation exercises. This is demonstrated by using massively parallel processing of the data vectorization for use in scalable machine learning. In addition, we show the utility of ensemble machine learning modeling for highly effective predictions of both goals of this challenge. This is demonstrated by the use of multiple machine learning algorithms combined with various scaling and normalization preprocessing steps. Then, using a voting ensemble, multiple models are combined to generate a final model prediction.

Ensemble machine learning modeling for the prediction of artemisinin resistance in malaria

Resistance in malaria is a growing concern affecting many areas of Sub-Saharan Africa and Southeast Asia. Since the emergence of artemisinin resistance in the late 2000s in Cambodia, research into the underlying mechanisms has been underway. The 2019 Malaria Challenge posited the task of developing computational models that address important problems in advancing the fight against malaria. The first goal was to accurately predict artemisinin drug resistance levels of Plasmodium falciparum isolates, as quantified by the IC 50. The second goal was to predict the parasite clearance rate of malaria parasite isolates based on in vitro transcriptional profiles. In this work, we develop machine learning models using novel methods for transforming isolate data and handling the tens of thousands of variables that result from these data transformation exercises. This is demonstrated by using massively parallel processing of the data vectorization for use in scalable machine learning. In addition, we show the utility of ensemble machine learning modeling for highly effective predictions of both goals of this challenge. This is demonstrated by the use of multiple machine learning algorithms combined with various scaling and normalization preprocessing steps. Then, using a voting ensemble, multiple models are combined to generate a final model prediction.

Ensemble machine learning modeling for the prediction of artemisinin resistance in malaria

Resistance in malaria is a growing concern affecting many areas of Sub-Saharan Africa and Southeast Asia. Since the emergence of artemisinin resistance in the late 2000s in Cambodia, research into the underlying mechanisms has been underway. The 2019 Malaria Challenge posited the task of developing computational models that address important problems in advancing the fight against malaria. The first goal was to accurately predict artemisinin drug resistance levels of Plasmodium falciparum isolates, as quantified by the IC 50. The second goal was to predict the parasite clearance rate of malaria parasite isolates based on in vitro transcriptional profiles. In this work, we develop machine learning models using novel methods for transforming isolate data and handling the tens of thousands of variables that result from these data transformation exercises. This is demonstrated by using massively parallel processing of the data vectorization for use in scalable machine learning. In addition, we show the utility of ensemble machine learning modeling for highly effective predictions of both goals of this challenge. This is demonstrated by the use of multiple machine learning algorithms combined with various scaling and normalization preprocessing steps. Then, using a voting ensemble, multiple models are combined to generate a final model prediction.

Machine learning approaches and their current application in plant molecular biology: A systematic review

Machine learning (ML) is a field of artificial intelligence that has rapidly emerged in molecular biology, thus allowing the exploitation of Big Data concepts in plant genomics. In this context, the main challenges are given in terms of how to analyze massive datasets and extract new knowledge in all levels of cellular systems research. In summary, ML techniques allow complex interactions to be inferred in several biological systems. Despite its potential, ML has been underused due to complex computational algorithms and definition terms. Therefore, a systematic review to disentangle ML approaches is relevant for plant scientists and has been considered in this study. We presented the main steps for ML development (from data selection to evaluation of classification/prediction models) with a respective discussion approaching functional genomics mainly in terms of pathogen effector genes in plant immunity. Additionally, we also considered how to access public source databases under an ML framework towards advancing plant molecular biology and introduced novel powerful tools, such as deep learning.

Dynamic Metabolic Footprinting Reveals the Key Components of Metabolic Network in Yeast Saccharomyces cerevisiae

Metabolic footprinting offers a relatively easy approach to exploit the potentials of metabolomics for phenotypic characterization of microbial cells. To capture the highly dynamic nature of metabolites, we propose the use of dynamic metabolic footprinting instead of the traditional method which relies on analysis at a single time point. Using direct infusion-mass spectrometry (DI-MS), we could observe the dynamic metabolic footprinting in yeast S. cerevisiae BY4709 (wild type) cultured on 3 different C-sources (glucose, glycerol, and ethanol) and sampled along 10 time points with 5 biological replicates. In order to analyze the dynamic mass spectrometry data, we developed the novel analysis methods that allow us to perform correlation analysis to identify metabolites that significantly correlate over time during growth on the different carbon sources. Both positive and negative electrospray ionization (ESI) modes were performed to obtain the complete information about the metabolite content. Using sparse principal component analysis (Sparse PCA), we further identified those pairs of metabolites that significantly contribute to the separation. From the list of significant metabolite pairs, we reconstructed an interaction map that provides information of how different metabolic pathways have correlated patterns during growth on the different carbon sources.

An analysis of a 'community-driven' reconstruction of the human metabolic network

Following a strategy similar to that used in baker's yeast (Herrgård et al. Nat Biotechnol 26:1155-1160, 2008). A consensus yeast metabolic network obtained from a community approach to systems biology (Herrgård et al. 2008; Dobson et al. BMC Syst Biol 4:145, 2010). Further developments towards a genome-scale metabolic model of yeast (Dobson et al. 2010; Heavner et al. BMC Syst Biol 6:55, 2012). Yeast 5-an expanded reconstruction of the Saccharomyces cerevisiae metabolic network (Heavner et al. 2012) and in Salmonella typhimurium (Thiele et al. BMC Syst Biol 5:8, 2011). A community effort towards a knowledge-base and mathematical model of the human pathogen Salmonellatyphimurium LT2 (Thiele et al. 2011), a recent paper (Thiele et al. Nat Biotechnol 31:419-425, 2013). A community-driven global reconstruction of human metabolism (Thiele et al. 2013) described a much improved 'community consensus' reconstruction of the human metabolic network, called Recon 2, and the authors (that include the present ones) have made it freely available via a database at http://humanmetabolism.org/ and in SBML format at Biomodels (http://identifiers.org/biomodels.db/MODEL1109130000). This short analysis summarises the main findings, and suggests some approaches that will be able to exploit the availability of this model to advantage.

The promiscuous binding of pharmaceutical drugs and their transporter-mediated uptake into cells: what we (need to) know and how we can do so

A recent paper in this journal sought to counter evidence for the role of transport proteins in effecting drug uptake into cells, and questions that transporters can recognize drug molecules in addition to their endogenous substrates. However, there is abundant evidence that both drugs and proteins are highly promiscuous. Most proteins bind to many drugs and most drugs bind to multiple proteins (on average more than six), including transporters (mutations in these can determine resistance); most drugs are known to recognise at least one transporter. In this response, we alert readers to the relevant evidence that exists or is required. This needs to be acquired in cells that contain the relevant proteins, and we highlight an experimental system for simultaneous genome-wide assessment of carrier-mediated uptake in a eukaryotic cell (yeast).

Functional network construction in Arabidopsis using rule-based machine learning on large-scale data sets

The meta-analysis of large-scale postgenomics data sets within public databases promises to provide important novel biological knowledge. Statistical approaches including correlation analyses in coexpression studies of gene expression have emerged as tools to elucidate gene function using these data sets. Here, we present a powerful and novel alternative methodology to computationally identify functional relationships between genes from microarray data sets using rule-based machine learning. This approach, termed "coprediction," is based on the collective ability of groups of genes co-occurring within rules to accurately predict the developmental outcome of a biological system. We demonstrate the utility of coprediction as a powerful analytical tool using publicly available microarray data generated exclusively from Arabidopsis thaliana seeds to compute a functional gene interaction network, termed Seed Co-Prediction Network (SCoPNet). SCoPNet predicts functional associations between genes acting in the same developmental and signal transduction pathways irrespective of the similarity in their respective gene expression patterns. Using SCoPNet, we identified four novel regulators of seed germination (ALTERED SEED GERMINATION5, 6, 7, and 8), and predicted interactions at the level of transcript abundance between these novel and previously described factors influencing Arabidopsis seed germination. An online Web tool to query SCoPNet has been developed as a community resource to dissect seed biology and is available at http://www.vseed.nottingham.ac.uk/.

Towards a unifying, systems biology understanding of large-scale cellular death and destruction caused by poorly liganded iron: Parkinson's, Huntington's, Alzheimer's, prions, bactericides, chemical toxicology and others as examples

Exposure to a variety of toxins and/or infectious agents leads to disease, degeneration and death, often characterised by circumstances in which cells or tissues do not merely die and cease to function but may be more or less entirely obliterated. It is then legitimate to ask the question as to whether, despite the many kinds of agent involved, there may be at least some unifying mechanisms of such cell death and destruction. I summarise the evidence that in a great many cases, one underlying mechanism, providing major stresses of this type, entails continuing and autocatalytic production (based on positive feedback mechanisms) of hydroxyl radicals via Fenton chemistry involving poorly liganded iron, leading to cell death via apoptosis (probably including via pathways induced by changes in the NF-κB system). While every pathway is in some sense connected to every other one, I highlight the literature evidence suggesting that the degenerative effects of many diseases and toxicological insults converge on iron dysregulation. This highlights specifically the role of iron metabolism, and the detailed speciation of iron, in chemical and other toxicology, and has significant implications for the use of iron chelating substances (probably in partnership with appropriate anti-oxidants) as nutritional or therapeutic agents in inhibiting both the progression of these mainly degenerative diseases and the sequelae of both chronic and acute toxin exposure. The complexity of biochemical networks, especially those involving autocatalytic behaviour and positive feedbacks, means that multiple interventions (e.g. of iron chelators plus antioxidants) are likely to prove most effective. A variety of systems biology approaches, that I summarise, can predict both the mechanisms involved in these cell death pathways and the optimal sites of action for nutritional or pharmacological interventions.

The photographer and the greenhouse: how to analyse plant metabolomics data

INTRODUCTION

Plant metabolomics experiments yield large amounts of data, too much to be interpretable by eye. Multivariate data analyses are therefore essential to extract and visualise the information of interest.

OBJECTIVE

Because multivariate statistical methods may be remote from the expertise of many scientists working in the metabolomics field, this overview provides a step-by-step description of a multivariate data analysis, starting from the experiment and ending with the figures appearing in scientific journals.

METHODOLOGY

We developed a thought experiment that explores the relationship between the differences in nutrient levels and three plant developmental descriptors through photography of the greenhouse they grow in. Through this, multivariate data analysis, data preprocessing and model validation are illustrated. Finally some of the presented methods are illustrated by the analysis of a plant metabolomics dataset.

CONCLUSION

This paper will familiarize non-specialised researchers with the main concepts in multivariate data analysis and allow them to develop and evaluate metabolomic data analyses more critically.

A metabolomic analysis of the toxicity of Aconitum sp. alkaloids in rats using gas chromatography/mass spectrometry

A metabolomic investigation of intoxication with Aconitum sp. alkaloids was carried out. Plasma obtained from Wistar rats administered these alkaloids was analyzed using gas chromatography/time-of-flight mass spectrometry. Samples were analyzed following protein precipitation, liquid-liquid extraction, and derivatization. Thirty-six metabolites were identified among the detected compounds. Subsequent data analysis using the t test and principal component analysis revealed metabolic differences between the control rats and treated groups as well as between the groups of rats administered different alkaloids. Twenty-seven metabolites were significantly different between plasma samples from the controls and treated groups. The significant decreased plasma levels of glutamine and creatinine in all treated groups suggested impaired heart and muscle function caused by alkaloids. The plasma levels of 22 metabolites in the hypaconitine group were significantly decreased. In contrast, only 8 and 13 metabolites were observed with significantly decreased levels in the aconitine and mesaconitine groups, respectively.These results indicated that Aconitum sp. alkaloids can cause metabolic disorders in rats. The toxicity and corresponding mechanism of hypaconitine was different from those of aconitine and mesaconitine, based on the differences of perturbed metabolic patterns between groups.

Preprocessing, classification modeling and feature selection using flow injection electrospray mass spectrometry metabolite fingerprint data

Metabolome analysis by flow injection electrospray mass spectrometry (FIE-MS) fingerprinting generates measurements relating to large numbers of m/z signals. Such data sets often exhibit high variance with a paucity of replicates, thus providing a challenge for data mining. We describe data preprocessing and modeling methods that have proved reliable in projects involving samples from a range of organisms. The protocols interact with software resources specifically for metabolomics provided in a Web-accessible data analysis package FIEmspro (http://users.aber.ac.uk/jhd) written in the R environment and requiring a moderate knowledge of R command-line usage. Specific emphasis is placed on describing the outcome of modeling experiments using FIE-MS data that require further preprocessing to improve quality. The salient features of both poor and robust (i.e., highly generalizable) multivariate models are outlined together with advice on validating classifiers and avoiding false discovery when seeking explanatory variables.

Metabolomic technologies and their application to the study of plants and plant-host interactions

Metabolomics is perhaps the ultimate level of post-genomic analysis as it can reveal changes in metabolite fluxes that are controlled by only minor changes within gene expression measured using transcriptomics and/or by analysing the proteome that elucidates post-translational control over enzyme activity. Metabolic change is a major feature of plant genetic modification and plant interactions with pathogens, pests, and their environment. In the assessment of genetically modified plant tissues, metabolomics has been used extensively to explore by-products resulting from transgene expression and scenarios of substantial equivalence. Many studies have concentrated on the physiological development of plant tissues as well as on the stress responses involved in heat shock or treatment with stress-eliciting molecules such as methyl jasmonic acid, yeast elicitor or bacterial lipopolysaccharide. Plant-host interactions represent one of the most biochemically complex and challenging scenarios that are currently being assessed by metabolomic approaches. For example, the mixtures of pathogen-colonised and non-challenged plant cells represent an extremely heterogeneous and biochemically rich sample; there is also the further complication of identifying which metabolites are derived from the plant host and which are from the interacting pathogen. This review will present an overview of the analytical instrumentation currently applied to plant metabolomic analysis, literature within the field will be reviewed paying particular regard to studies based on plant-host interactions and finally the future prospects on the metabolomic analysis of plants and plant-host interactions will be discussed.

Identification of hexose diastereomers by means of tandem mass spectrometry of oxocarbenium ions followed by neural networks analysis

The ion abundances in the tandem mass spectra of oxocarbenium ions generated from aldohexoses and ketohexoses under electrospray ionization conditions depend on their stereochemistry. The ratios of the abundances of two of the major ions m/z 85 and 91 are different in the aldohexoses and the ketohexoses. Principal component analysis (PCA) allowed visual differentiation of the hexose isomers. However, identification of an individual hexose was difficult. The data were subjected to neural network analysis (NNA) using a multilayer perceptron (MLP) model. The model was first trained using the data from the 11 hexoses and then used to identify the hexoses using fresh data acquired later. There was 100% identification of the various hexoses. The hexose units in methyl glucoside, mannoside and galactoside could also be identified accurately. The method can therefore be used for the characterization of hexose units from monosaccharides and glycosides.

High-throughput, nontargeted metabolite fingerprinting using nominal mass flow injection electrospray mass spectrometry

Flow injection electrospray-mass spectrometry (FIE-MS) is finding utility as a first-pass metabolite fingerprinting tool in many research fields. We provide a protocol that has proved reliable in large-scale research projects involving diverse sample matrices originating from plants, microbes and mammalian biofluids. Using Brachypodium leaf material as an example matrix all steps are summarized from sample extraction to data quality assessment. Alternative procedures for dealing with other common matrices such as bloods and urine are included. With little sample pretreatment, no chromatography and instrument cycle times of <5 min it is feasible to analyze >1,000 samples per week. Analysis of a typical batch of 240 samples (including first-pass data analysis) can be accomplished within 48 h.

Proteomic applications in ecotoxicology

Within the growing body of proteomics studies, issues addressing problems of ecotoxicology are on the rise. Generally speaking, ecotoxicology uses quantitative expression changes of distinct proteins known to be involved in toxicological responses as biomarkers. Unlike these directed approaches, proteomics examines how multiple expression changes are associated with a contamination that is suspected to be detrimental. Consequently, proteins involved in toxicological responses that have not been described previously may be revealed. Following identification of key proteins indicating exposure or effect, proteomics can potentially be employed in environmental risk assessment. To this end, bioinformatics may unveil protein patterns specific to an environmental stress that would constitute a classifier able to distinguish an exposure from a control state. The combined use of sets of marker proteins associated with a given pollution impact may prove to be more reliable, as they are based not only on a few unique markers which are measured independently, but reflect the complexity of a toxicological response. Such a proteomic pattern might also integrate some of the already established biomarkers of environmental toxicity. Proteomics applications in ecotoxicology may also comprise functional examination of known classes of proteins, such as glutathione transferases or metallothioneins, to elucidate their toxicological responses.

Neural network analyses of infrared spectra for classifying cell wall architectures

About 10% of plant genomes are devoted to cell wall biogenesis. Our goal is to establish methodologies that identify and classify cell wall phenotypes of mutants on a genome-wide scale. Toward this goal, we have used a model system, the elongating maize (Zea mays) coleoptile system, in which cell wall changes are well characterized, to develop a paradigm for classification of a comprehensive range of cell wall architectures altered during development, by environmental perturbation, or by mutation. Dynamic changes in cell walls of etiolated maize coleoptiles, sampled at one-half-d intervals of growth, were analyzed by chemical and enzymatic assays and Fourier transform infrared spectroscopy. The primary walls of grasses are composed of cellulose microfibrils, glucuronoarabinoxylans, and mixed-linkage (1 --> 3),(1 --> 4)-beta-D-glucans, together with smaller amounts of glucomannans, xyloglucans, pectins, and a network of polyphenolic substances. During coleoptile development, changes in cell wall composition included a transient appearance of the (1 --> 3),(1 --> 4)-beta-D-glucans, a gradual loss of arabinose from glucuronoarabinoxylans, and an increase in the relative proportion of cellulose. Infrared spectra reflected these dynamic changes in composition. Although infrared spectra of walls from embryonic, elongating, and senescent coleoptiles were broadly discriminated from each other by exploratory principal components analysis, neural network algorithms (both genetic and Kohonen) could correctly classify infrared spectra from cell walls harvested from individuals differing at one-half-d interval of growth. We tested the predictive capabilities of the model with a maize inbred line, Wisconsin 22, and found it to be accurate in classifying cell walls representing developmental stage. The ability of artificial neural networks to classify infrared spectra from cell walls provides a means to identify many possible classes of cell wall phenotypes. This classification can be broadened to phenotypes resulting from mutations in genes encoding proteins for which a function is yet to be described.

Metabolomics of a superorganism

The human can be thought of as a human-microbe hybrid, and the health of this superorganism will be affected by intrinsic properties such as human genetics, diurnal cycles, and age and by extrinsic factors such as lifestyle choices (food and drink, drug intake) and the acquisition of a stable "healthy" gut microflora (the so-called microbiome). Alterations in this superorganism will be manifest in the metabolite complement within its serum and urine samples. The unraveling of this metabolic compartmentalization in this complex ecosystem will certainly be a challenge for systems biology and necessary for defining human health at the molecular level. Within the systems biology framework, functional analyses at the level of gene expression (transcriptomics), protein translation (proteomics), and, more recently, the metabolite network (metabolomics) have become increasingly popular. Metabolomics experiments aim to quantify all metabolites in a cellular system (cell or tissue) under defined states and at different time points so that the dynamics of any biotic, abiotic, or genetic perturbation can be accurately assessed. This article provides an overview of metabolomics and discusses how data are generated and analyzed within a systems biology framework. The role of metabolomics in nutrigenomics is also discussed, as are the concepts of the human being a superorganism and the complexities required to be overcome to understand human health and disease.

Metabolomics: current technologies and future trends

The ability to sequence whole genomes has taught us that our knowledge with respect to gene function is rather limited with typically 30-40% of open reading frames having no known function. Thus, within the life sciences there is a need for determination of the biological function of these so-called orphan genes, some of which may be molecular targets for therapeutic intervention. The search for specific mRNA, proteins, or metabolites that can serve as diagnostic markers has also increased, as has the fact that these biomarkers may be useful in following and predicting disease progression or response to therapy. Functional analyses have become increasingly popular. They include investigations at the level of gene expression (transcriptomics), protein translation (proteomics) and more recently the metabolite network (metabolomics). This article provides an overview of metabolomics and discusses its complementary role with transcriptomics and proteomics, and within system biology. It highlights how metabolome analyses are conducted and how the highly complex data that are generated are analysed. Non-invasive footprinting analysis is also discussed as this has many applications to in vitro cell systems. Finally, for studying biotic or abiotic stresses on animals, plants or microbes, we believe that metabolomics could very easily be applied to large populations, because this approach tends to be of higher throughput and generally lower cost than transcriptomics and proteomics, whilst also providing indications of which area of metabolism may be affected by external perturbation.

Predicting interpretability of metabolome models based on behavior, putative identity, and biological relevance of explanatory signals

Powerful algorithms are required to deal with the dimensionality of metabolomics data. Although many achieve high classification accuracy, the models they generate have limited value unless it can be demonstrated that they are reproducible and statistically relevant to the biological problem under investigation. Random forest (RF) generates models, without any requirement for dimensionality reduction or feature selection, in which individual variables are ranked for significance and displayed in an explicit manner. In metabolome fingerprinting by mass spectrometry, each metabolite can be represented by signals at several m/z. Exploiting a prior understanding of expected biochemical differences between sample classes, we aimed to develop meaningful metrics relevant to the significance both of the overall RF model and individual, potentially explanatory, signals. Pair-wise comparison of related plant genotypes with strong phenotypic differences demonstrated that robust models are not only reproducible but also logically structured, highlighting correlated m/z derived from just a small number of explanatory metabolites reflecting the biological differences between sample classes. RF models were also generated by using groupings of samples known to be increasingly phenotypically similar. Although classification accuracy was often reasonable, we demonstrated reproducibly in both Arabidopsis and potato a performance threshold based on margin statistics beyond which such models showed little structure indicative of either generalizability or further biological interpretability. In a multiclass problem using 25 Arabidopsis genotypes, despite the complicating effects of ecotype background and secondary metabolome perturbations common to several mutations, the ranking of metabolome signals by RF provided scope for deeper interpretability.

Data processing and classification analysis of proteomic changes: a case study of oil pollution in the mussel, Mytilus edulis

BACKGROUND

Proteomics may help to detect subtle pollution-related changes, such as responses to mixture pollution at low concentrations, where clear signs of toxicity are absent. The challenges associated with the analysis of large-scale multivariate proteomic datasets have been widely discussed in medical research and biomarker discovery. This concept has been introduced to ecotoxicology only recently, so data processing and classification analysis need to be refined before they can be readily applied in biomarker discovery and monitoring studies.

RESULTS

Data sets obtained from a case study of oil pollution in the Blue mussel were investigated for differential protein expression by retentate chromatography-mass spectrometry and decision tree classification. Different tissues and different settings were used to evaluate classifiers towards their discriminatory power. It was found that, due the intrinsic variability of the data sets, reliable classification of unknown samples could only be achieved on a broad statistical basis (n > 60) with the observed expression changes comprising high statistical significance and sufficient amplitude. The application of stringent criteria to guard against overfitting of the models eventually allowed satisfactory classification for only one of the investigated data sets and settings.

CONCLUSION

Machine learning techniques provide a promising approach to process and extract informative expression signatures from high-dimensional mass-spectrometry data. Even though characterisation of the proteins forming the expression signatures would be ideal, knowledge of the specific proteins is not mandatory for effective class discrimination. This may constitute a new biomarker approach in ecotoxicology, where working with organisms, which do not have sequenced genomes render protein identification by database searching problematic. However, data processing has to be critically evaluated and statistical constraints have to be considered before supervised classification algorithms are employed.

A bioinformatician's view of the metabolome

The study of a collection of metabolites as a whole (metabolome), as opposed to isolated small molecules, is a fast-growing field promising to take us one step further towards understanding cell biology, and relating the genetic capabilities of an organism to its observed phenotype. The new sciences of metabolomics and metabonomics can exploit a variety of existing experimental and computational methods, but they also require new technology that can deal with both the amount and the diversity of the data relating to the rich world of metabolites. More specifically, the collaboration between bioinformaticians and chemoinformaticians promises to advance our view of cognate molecules, by shedding light on their atomic structure and properties. Modelling of the interactions of metabolites with other entities in the cell, and eventually complete modelling of reaction pathways will be essential for analysis of the experimental data, and prediction of an organism's response to environmental challenges.

Metabolic fingerprinting in disease diagnosis: biomedical applications of infrared and Raman spectroscopy

The ability to diagnose the early onset of disease, rapidly, non-invasively and unequivocally has multiple benefits. These include the early intervention of therapeutic strategies leading to a reduction in morbidity and mortality, and the releasing of economic resources within overburdened health care systems. Some of the routine clinical tests currently in use are known to be unsuitable or unreliable. In addition, these often rely on single disease markers which are inappropriate when multiple factors are involved. Many diseases are a result of metabolic disorders, therefore it is logical to measure metabolism directly. One of the strategies employed by the emergent science of metabolomics is metabolic fingerprinting; which involves rapid, high-throughput global analysis to discriminate between samples of different biological status or origin. This review focuses on a selective number of recent studies where metabolic fingerprinting has been forwarded as a potential tool for disease diagnosis using infrared and Raman spectroscopies.

Theodor Bücher Lecture. Metabolomics, modelling and machine learning in systems biology - towards an understanding of the languages of cells. Delivered on 3 July 2005 at the 30th FEBS Congress and the 9th IUBMB conference in Budapest

The newly emerging field of systems biology involves a judicious interplay between high-throughput 'wet' experimentation, computational modelling and technology development, coupled to the world of ideas and theory. This interplay involves iterative cycles, such that systems biology is not at all confined to hypothesis-dependent studies, with intelligent, principled, hypothesis-generating studies being of high importance and consequently very far from aimless fishing expeditions. I seek to illustrate each of these facets. Novel technology development in metabolomics can increase substantially the dynamic range and number of metabolites that one can detect, and these can be exploited as disease markers and in the consequent and principled generation of hypotheses that are consistent with the data and achieve this in a value-free manner. Much of classical biochemistry and signalling pathway analysis has concentrated on the analyses of changes in the concentrations of intermediates, with 'local' equations - such as that of Michaelis and Menten v=(Vmax x S)/(S+K m) - that describe individual steps being based solely on the instantaneous values of these concentrations. Recent work using single cells (that are not subject to the intellectually unsupportable averaging of the variable displayed by heterogeneous cells possessing nonlinear kinetics) has led to the recognition that some protein signalling pathways may encode their signals not (just) as concentrations (AM or amplitude-modulated in a radio analogy) but via changes in the dynamics of those concentrations (the signals are FM or frequency-modulated). This contributes in principle to a straightforward solution of the crosstalk problem, leads to a profound reassessment of how to understand the downstream effects of dynamic changes in the concentrations of elements in these pathways, and stresses the role of signal processing (and not merely the intermediates) in biological signalling. It is this signal processing that lies at the heart of understanding the languages of cells. The resolution of many of the modern and postgenomic problems of biochemistry requires the development of a myriad of new technologies (and maybe a new culture), and thus regular input from the physical sciences, engineering, mathematics and computer science. One solution, that we are adopting in the Manchester Interdisciplinary Biocentre (http://www.mib.ac.uk/) and the Manchester Centre for Integrative Systems Biology (http://www.mcisb.org/), is thus to colocate individuals with the necessary combinations of skills. Novel disciplines that require such an integrative approach continue to emerge. These include fields such as chemical genomics, synthetic biology, distributed computational environments for biological data and modelling, single cell diagnostics/bionanotechnology, and computational linguistics/text mining.

Integrating metabolomics into a systems biology framework to exploit metabolic complexity: strategies and applications in microorganisms

As an important functional genomic tool, metabolomics has been illustrated in detail in recent years, especially in plant science. However, the microbial category also has the potential to benefit from integration of metabolomics into system frameworks. In this article, we first examine the concepts and brief history of metabolomics. Next, we summarize metabolomic research processes and analytical platforms in strain improvements. The application cases of metabolomics in microorganisms answer what the metabolomics can do in strain improvements. The position of metabolomics in this systems biology framework and the real cases of integrating metabolomics into a system framework to explore the microbial metabolic complexity are also illustrated in this paper.

Hierarchical metabolomics demonstrates substantial compositional similarity between genetically modified and conventional potato crops

There is current debate whether genetically modified (GM) plants might contain unexpected, potentially undesirable changes in overall metabolite composition. However, appropriate analytical technology and acceptable metrics of compositional similarity require development. We describe a comprehensive comparison of total metabolites in field-grown GM and conventional potato tubers using a hierarchical approach initiating with rapid metabolome "fingerprinting" to guide more detailed profiling of metabolites where significant differences are suspected. Central to this strategy are data analysis procedures able to generate validated, reproducible metrics of comparison from complex metabolome data. We show that, apart from targeted changes, these GM potatoes in this study appear substantially equivalent to traditional cultivars.

Mass spectrometry in metabolome analysis

In the post-genomic era, increasing efforts have been made to describe the relationship between the genome and the phenotype in cells and organisms. It has become clear that even a complete understanding of the state of the genes, messages, and proteins in a living system does not reveal its phenotype. Therefore, researchers have started to study the metabolome (or the metabolic complement of functional genomics). Within this context, mass spectrometry (MS) has increasingly occupied a central position in the methodologies developed for determination of the metabolic state. This review is mainly focused on the status of MS in the metabolome field, trying to direct the reader to the main approaches for analysis of metabolites, reviewing basic methodologies in sample preparation, and the most recent MS techniques introduced. Apart from the description of the different methods, this review will try to state a general comparison between the several different techniques that involve MS and metabolite analysis, and will highlight their limitations and preferred applicability.

Strategy for metabonomics research based on high-performance liquid chromatography and liquid chromatography coupled with tandem mass spectrometry

Metabonomics, the study of metabolites and their roles in various disease states, is a novel methodology arising from the post-genomics era. This methodology has been applied in many fields. Current metabonomic practice has relied on mass spectrometry (MS), gas chromatography-mass spectrometry (GC-MS), and nuclear magnetic resonance (NMR) to analyze metabolites. In this study, a strategy was developed for applying high-performance liquid chromatography (HPLC) and LC-MS-MS to metabonomics research. One of the key problems to be solved in this strategy is to match the peaks between the chromatograms. A peak alignment algorithm has been developed to match the chromatograms before the pattern recognition. As an application example, the strategy described above was applied to metabonomics research on liver diseases, and the false-positive result of live cancer diagnosis from the hepatocirrhosis and hepatitis diseases was effectively reduced to 7.40%. Based on the pattern recognition, several potential biomarkers were found and further identified by the following LC-MS-MS experiments. The structures of eight potential biomarkers were given for distinguishing the liver cancer from the hepatocirrhosis and hepatitis diseases.

Metabolic footprinting and systems biology: the medium is the message

One element of classical systems analysis treats a system as a black or grey box, the inner structure and behaviour of which can be analysed and modelled by varying an internal or external condition, probing it from outside and studying the effect of the variation on the external observables. The result is an understanding of the inner make-up and workings of the system. The equivalent of this in biology is to observe what a cell or system excretes under controlled conditions - the 'metabolic footprint' or exometabolome - as this is readily and accurately measurable. Here, we review the principles, experimental approaches and scientific outcomes that have been obtained with this useful and convenient strategy.

Metabolomics, machine learning and modelling: towards an understanding of the language of cells

In answering the question ‘Systems Biology – will it work?’ (which it self-evidently has already), it is appropriate to highlight advances in philosophy, in new technique development and in novel findings. In terms of philosophy, we see that systems biology involves an iterative interplay between linked activities – for instance, between theory and experiment, between induction and deduction and between measurements of parameters and variables – with more emphasis than has perhaps been common now being focused on the first in each of these pairs. In technique development, we highlight closed loop machine learning and its use in the optimization of scientific instrumentation, and the ability to effect high-quality and quasi-continuous optical images of cells. This leads to many important and novel findings. In the first case, these may involve new biomarkers for disease, whereas in the second case, we have determined that many biological signals may be frequency-rather than amplitude-encoded. This leads to a very different view of how signalling ‘works’ (equations such as that of Michaelis and Menten which use only amplitudes, i.e. concentrations, are inadequate descriptors), lays emphasis on the signal processing network elements that lie ‘downstream’ of what are traditionally considered the signals, and allows one simply to understand how cross-talk may be avoided between pathways which nevertheless use common signalling elements. The language of cells is much richer than we had supposed, and we are now well placed to decode it.

Reverse engineering of biochemical equations from time-course data by means of genetic programming

Increased research aimed at simulating biological systems requires sophisticated parameter estimation methods. All current approaches, including genetic algorithms, need pre-existing equations to be functional. A generalized approach to predict not only parameters but also biochemical equations from only observable time-course information must be developed and a computational method to generate arbitrary equations without knowledge of biochemical reaction mechanisms must be developed. We present a technique to predict an equation using genetic programming. Our technique can search topology and numerical parameters of mathematical expression simultaneously. To improve the search ability of numeric constants, we added numeric mutation to the conventional procedure. As case studies, we predicted two equations of enzyme-catalyzed reactions regarding adenylate kinase and phosphofructokinase. Our numerical experimental results showed that our approach could obtain correct topology and parameters that were close to the originals. The mean errors between given and simulation-predicted time-courses were 1.6 x 10(-5)% and 2.0 x 10(-3)%, respectively. Our equation prediction approach can be applied to identify metabolic reactions from observable time-courses.

Discrimination of modes of action of antifungal substances by use of metabolic footprinting

Diploid cells of Saccharomyces cerevisiae were grown under controlled conditions with a Bioscreen instrument, which permitted the essentially continuous registration of their growth via optical density measurements. Some cultures were exposed to concentrations of a number of antifungal substances with different targets or modes of action (sterol biosynthesis, respiratory chain, amino acid synthesis, and the uncoupler). Culture supernatants were taken and analyzed for their "metabolic footprints" by using direct-injection mass spectrometry. Discriminant function analysis and hierarchical cluster analysis allowed these antifungal compounds to be distinguished and classified according to their modes of action. Genetic programming, a rule-evolving machine learning strategy, allowed respiratory inhibitors to be discriminated from others by using just two masses. Metabolic footprinting thus represents a rapid, convenient, and information-rich method for classifying the modes of action of antifungal substances.

Application of high-throughput Fourier-transform infrared spectroscopy in toxicology studies: contribution to a study on the development of an animal model for idiosyncratic toxicity

An evaluation of high-throughput Fourier-transform infrared spectroscopy (FT-IR) as a technology that could support a "metabonomics" component in toxicological studies of drug candidates is presented. The hypothesis tested in this study was that FT-IR had sufficient resolving power to discriminate between urine collected from control rat populations and rats subjected to treatment with a potent inflammatory agent, bacterial lipopolysaccharide (LPS). It was also hypothesized that co-administration of LPS with ranitidine, a drug associated with reports of idiosyncratic susceptibility, would induce hepatotoxicity in rats and that this could be detected non-invasively by an FT-IR-based metabonomics approach. The co-administration of LPS with "idiosyncratic" drugs represents an attempt to develop a predictive model of idiosyncratic toxicity and FT-IR is used herein to support characterization of this model. FT-IR spectra are high dimensional and the use of genetic programming to identify spectral sub-regions that most contribute to discrimination is demonstrated. FT-IR is rapid, reagentless, highly reproducible and inexpensive. Results from this pilot study indicate it could be extended to routine applications in toxicology and to supporting characterization of a new animal model for idiosyncratic susceptibility.

Investigating plant-plant interference by metabolic fingerprinting

New analytical developments in post-genomic technologies are being introduced to the field of plant ecology. FT-IR fingerprinting coupled with chemometrics via cluster analysis is proposed as a tool for correlating global metabolic changes with abiotic or biotic perturbation and/or interactions. The current study concentrates on detecting chemical responses by inter-species competition between a monocotyledon Brachypodium distachyion and a dicotyledon Arabidopsis thaliana. Growth analysis of 42 days old plants showed differences in both species under competition. Clear changes in the FT-IR metabolic fingerprints of B. distachyion in competition with A. thaliana were observed, whilst there were no apparent chemical differences in the A. thaliana plant tissues. This study demonstrates the power of this approach in detecting changes in the global metabolic profiles of plants in response to biotic interactions, and we believe FT-IR is appropriate for rapid screening (10 s per sample) prior to targeted metabolite analyses.

High-throughput classification of yeast mutants for functional genomics using metabolic footprinting

Many technologies have been developed to help explain the function of genes discovered by systematic genome sequencing. At present, transcriptome and proteome studies dominate large-scale functional analysis strategies. Yet the metabolome, because it is 'downstream', should show greater effects of genetic or physiological changes and thus should be much closer to the phenotype of the organism. We earlier presented a functional analysis strategy that used metabolic fingerprinting to reveal the phenotype of silent mutations of yeast genes. However, this is difficult to scale up for high-throughput screening. Here we present an alternative that has the required throughput (2 min per sample). This 'metabolic footprinting' approach recognizes the significance of 'overflow metabolism' in appropriate media. Measuring intracellular metabolites is time-consuming and subject to technical difficulties caused by the rapid turnover of intracellular metabolites and the need to quench metabolism and separate metabolites from the extracellular space. We therefore focused instead on direct, noninvasive, mass spectrometric monitoring of extracellular metabolites in spent culture medium. Metabolic footprinting can distinguish between different physiological states of wild-type yeast and between yeast single-gene deletion mutants even from related areas of metabolism. By using appropriate clustering and machine learning techniques, the latter based on genetic programming, we show that metabolic footprinting is an effective method to classify 'unknown' mutants by genetic defect.

An optimized protocol for metabolome analysis in yeast using direct infusion electrospray mass spectrometry

A method for the global analysis of yeast intracellular metabolites, based on electrospray mass spectrometry (ES-MS), has been developed. This has involved the optimization of methods for quenching metabolism in Saccharomyces cerevisiae and extracting the metabolites for analysis by positive-ion electrospray mass spectrometry. The influence of cultivation conditions, sampling, quenching and extraction conditions, concentration step, and storage have all been studied and adapted to allow direct infusion of samples into the mass spectrometer and the acquisition of metabolic profiles with simultaneous detection of more than 25 intracellular metabolites. The method, which can be applied to other micro-organisms and biological systems, may be used for comparative analysis and screening of metabolite profiles of yeast strains and mutants under controlled conditions in order to elucidate gene function via metabolomics. Examples of the application of this analytical strategy to specific yeast strains and single-ORF yeast deletion mutants generated through the EUROFAN programme are presented.

Chemometric discrimination of unfractionated plant extracts analyzed by electrospray mass spectrometry

Metabolic fingerprints were obtained from unfractionated Pharbitis nil leaf sap samples by direct infusion into an electrospray ionization mass spectrometer. Analyses took less than 30 s per sample and yielded complex mass spectra. Various chemometric methods, including discriminant function analysis and the machine-learning methods of artificial neural networks and genetic programming, could discriminate the metabolic fingerprints of plants subjected to different photoperiod treatments. This rapid automated analytical procedure could find use in a variety of phytochemical applications requiring high sample throughput.

Metabolomics in systems biology

The primary aim of "omic" technologies is the nontargeted identification of all gene products (transcripts, proteins, and metabolites) present in a specific biological sample. By their nature, these technologies reveal unexpected properties of biological systems. A second and more challenging aspect of omic technologies is the refined analysis of quantitative dynamics in biological systems. For metabolomics, gas and liquid chromatography coupled to mass spectrometry are well suited for coping with high sample numbers in reliable measurement times with respect to both technical accuracy and the identification and quantitation of small-molecular-weight metabolites. This potential is a prerequisite for the analysis of dynamic systems. Thus, metabolomics is a key technology for systems biology. The aim of this review is to (a) provide an in-depth overview about metabolomic technology, (b) explore how metabolomic networks can be connected to the underlying reaction pathway structure, and (c) discuss the need to investigate integrative biochemical networks.

Genotype-phenotype mapping: genes as computer programs

The effects of genes on phenotype are mediated by processes that are typically unknown but whose determination is desirable. The conversion from gene to phenotype is not a simple function of individual genes, but involves the complex interactions of many genes; it is what is known as a nonlinear mapping problem. A computational method called genetic programming allows the representation of candidate nonlinear mappings in several possible trees. To find the best model, the trees are 'evolved' by processes akin to mutation and recombination, and the trees that more closely represent the actual data are preferentially selected. The result is an improved tree of rules that represent the nonlinear mapping directly. In this way, the encoding of cellular and higher-order activities by genes is seen as directly analogous to computer programs. This analogy is of utility in biological genetics and in problems of genotype-phenotype mapping.

Rapid and quantitative detection of the microbial spoilage of meat by fourier transform infrared spectroscopy and machine learning

Fourier transform infrared (FT-IR) spectroscopy is a rapid, noninvasive technique with considerable potential for application in the food and related industries. We show here that this technique can be used directly on the surface of food to produce biochemically interpretable "fingerprints." Spoilage in meat is the result of decomposition and the formation of metabolites caused by the growth and enzymatic activity of microorganisms. FT-IR was exploited to measure biochemical changes within the meat substrate, enhancing and accelerating the detection of microbial spoilage. Chicken breasts were purchased from a national retailer, comminuted for 10 s, and left to spoil at room temperature for 24 h. Every hour, FT-IR measurements were taken directly from the meat surface using attenuated total reflectance, and the total viable counts were obtained by classical plating methods. Quantitative interpretation of FT-IR spectra was possible using partial least-squares regression and allowed accurate estimates of bacterial loads to be calculated directly from the meat surface in 60 s. Genetic programming was used to derive rules showing that at levels of 10(7) bacteria.g(-1) the main biochemical indicator of spoilage was the onset of proteolysis. Thus, using FT-IR we were able to acquire a metabolic snapshot and quantify, noninvasively, the microbial loads of food samples accurately and rapidly in 60 s, directly from the sample surface. We believe this approach will aid in the Hazard Analysis Critical Control Point process for the assessment of the microbiological safety of food at the production, processing, manufacturing, packaging, and storage levels.