Search

This chapter provides an overview of emerging bioinformatics methods for the biomarker discovery process and medical decision support. It introduces study design consideration and bioanalytic concepts for generating biomedical data, followed by various data mining and information retrieval procedures such as feature selection, classification as well as statistical and clinical validation. The reviewed methods are illustrated by real examples from preclinical and clinical studies, and the application in medical decision making is discussed. This chapter is anticipated to address to those with a bioinformatics background as well as biomedical researchers who are interested in the application of computational methods in biomarker discovery and medical decision making.

Motivation: During the Bavarian newborn screening programme all newborns have been tested for about 20 inherited metabolic disorders. Owing to the amount and complexity of the generated experimental data, machine learning techniques provide a promising approach to investigate novel patterns in high-dimensional metabolic data which form the source for constructing classification rules with high discriminatory power. Results: Six machine learning techniques have been investigated for their classification accuracy focusing on two metabolic disorders, phenylketo nuria (PKU) and medium-chain acyl-CoA dehydrogenase deficiency (MCADD). Logistic regression analysis led to superior classification rules (sensitivity >96.8%, specificity >99.98%) compared to all investigated algorithms. Including novel constellations of metabolites into the models, the positive predictive value could be strongly increased (PKU 71.9% versus 16.2%, MCADD 88.4% versus 54.6% compared to the established diagnostic markers). Our results clearly prove that the mined data confirm the known and indicate some novel metabolic patterns which may contribute to a better understanding of newborn metabolism. Availability: WEKA machine learning package: www.cs.waikato.ac.nz/~ml/weka and statistical software package ADE-4: http://pbil.univ-lyon1.fr/ADE-4

The relationship between hydrological connectivity and species diversity patterns (alpha and beta diversity) of macrophytes, molluscs, odonates and amphibians was investigated in a semi-natural floodplain segment in the ‘Alluvial Zone National Park’ of the Danube River in Austria. Based on environmental variables, we distinguished four major channel types (inflow channel, parapotamal, plesiopotamal and palaeopotamal) that reflected a lateral connectivity gradient. In addition, a longitudinal environmental gradient along the parapotamal channel was found. Connectivity, rather than the surface area of individual floodplain water bodies, explained local species richness. Species diversity patterns varied among taxa: the highest species richness values for molluscs occurred in the parapotamal channels, for odonates in the para- and plesiopotamal channels, for macrophytes in the plesiopotamal channels and for amphibians in the palaeopotamal channels. Within the parapotamal channels, the species richness of odonates and amphibians increased moving upstream. Beta diversity displayed an almost inverse relationship with alpha diversity, with highest average values in isolated and fragmented floodplain channels. Habitat fragmentation favoured the beta diversity of most groups, although connectivity favoured the beta diversity of amphibians. The highest proportion of endangered species (mainly rheophilic forms) was found in the parapotamal channels. It is concluded that preservation of the high diversity of this alluvial flood plain would be more fully realised by reconstitution of fluvial dynamics and the associated connectivity gradients, rather than by restoration strategies for individual groups or endangered species. Copyright © 1999 John Wiley & Sons, Ltd.

The malignant clone in acute myeloid leukemia (AML) is organized hierarchically with more mature cells programmed to undergo natural apoptosis after a variable number of cell divisions and immature primitive cells that have self-renewal and NOD/SCID mouse-repopulating capacity. In most types of AML, leukemic stem cells supposedly reside within the CD34+/CD38− fraction of the leukemic clone. Because of their long term leukemia-repopulating capacity, AML stem cells are a logic target of therapy. However, so far, little is known about the regulation of growth and survival of these cells. We examined the expression of cytokine receptors (SCFR/KIT, IL-3Rα, GM-CSFRα, IL-3/GM-CSFRβ, G-CSFR, M-CSFR, TGFβR/endoglin, EPOR, TPOR/MPL, FLT3, VEGFR/KDR) and of other molecular targets and markers (CD33, CD44, CD133) on CD34+/CD38− cells in patients with AML (n=30), and determined responses to cytokine-ligands, conventional antileukemic drugs (ARA-C, fludarabine) or targeted drugs (Gemtuzumab/Ozogamicin, GO = Mylotarg®) in these cells. Apoptosis in AML stem cells was analyzed by combined staining for surface markers and AnnexinV. In a group of patients, CD34+/CD38- cells were purified to homogeneity by cell sorting and examined for 3H-thymidine uptake. Unexpectedly, AML stem cells were found to display a highly variable pattern of cytokine receptors and cell surface targets. In fact, only the IL-3R and CD44 were expressed consistently on all AML stem cells in all donors tested. In most patients, at least a subset of AML stem cells also co-expressed the SCFR/KIT, G-CSFR, TGFβR, FLT3, CD33, and CD133. By contrast, AML stem cells in most donors were found to lack substantial amounts of the GM-CSFR, M-CSFR, EPOR, TPOR, and VEGFR/KDR. When cultured in RPMI-1640 medium plus 10% FCS, about 5–15% of the CD34+/CD38- cells and about 10–40% of the more mature CD34+ AML cells were found to undergo spontaneous apoptosis within 48 hours. Spontaneous apoptosis was prevented by exposure to SCF, IL-3, or G-CSF, but not by exposure to EPO. ARA-C (0.5–5 μg/ml), fludarabine (0.1–5 μg/ml), and GO/Mylotarg® (1 μg/ml) were found to promote apoptosis in CD34+/CD38− cells and in more mature AML cells in all donors tested. The effects of ARA-C and fludarabine on AML cells were found to be dose-dependent. Moreover, we were able to show that ARA-C (0.5 μg/ml) and fludarabine (0.1 μg/ml) cooperate with each other in producing apoptosis in AML (stem) cells. 3H-thymidine uptake experiments performed on purified CD34+/CD38− AML stem cells confirmed the growth-inhibitory effects of these drugs. In particular, ARA-C and fludarabine were found to inhibit the cytokine-induced 3H-thymidine uptake in sorted AML stem cells in all donors examined. In summary, our data show that multi-color flow cytometry and combined staining for surface markers and AnnexinV is a powerful approach to determine apoptosis-preventing effects of cytokines and apoptosis-inducing effects of anti-leukemic drugs in immature CD34+/CD38− AML progenitor cells. Using this assay, it should be possible to identify combinations of targeted and/or conventional drugs eliminating maximal numbers of leukemic stem cells in AML.

Computed tomography (CT) is considered an upcoming technology for the quantitative estimation of myocardial perfusion (MPF). Based on previous results that affirm the feasibility of estimating global MPF using a CT scanner in an animal model, this work evaluates the computation of MPF in patients by comparing two different mathematical approaches based on the indicator dilution method. The first, the so called upslope method, was already used in the previous study, the second uses a deconvolution model which possibly is more suitable in the evaluation of regional MPF measurements. A good correlation between both methods could be found and the quantitative results of both methods correspond well to myocardial perfusion values as described in the literature.

Die myokardiale Perfusion stellt einen wesentlichen Parameter der funktionellen. Determinanten des Herzens dar. Die Bestimmung dieses Parameters mittels direkter Mesverfahren in vivo am Menschen ist nicht vertretbar. So kommen verschiedene indirekte Mesmethoden (Fremdgas-, Thermodilution-, Dopplerkatheter- und radioaktive Techniken) zum Ensatz. Seit der Entwicklung der Electron-Beam-computertomographischen Technologie ist es moglich geworden, nicht nur die linskventrikularen Volumina zu bestimmen, sondern auch die myokardiale Perfusion in ml/100 g/min zu berechnen. Die Messung der linksventrikularen Myokardperfusion erfolgt hierbei im sogenannten Multi-Slice-Flow-Mode in “short axis view”. Hierbei wird der Untersuchungstisch mit dem Patienten 25° nach rechts und 15° nach kaudal gekippt. Nach einem Lokalisationsscan zur Bestimmung der Lage des linken Ventrikels wird der Tisch mit dem Patienten so positioniert, das sich die voraussichtliche Mitte des linken Ventrikels im Mesfeld der dritten und vierten CT-Schicht befindet. Nach intravenoser Kontrastmittelapplikation von 50 ml und einem Flow von 3 ml/s werden 13×6 CT-Schichten EKG-getriggert jeden zweiten oder dritten Herzschlag angefertigt. Auf der Befundkonsole werden die Regions of Interest so in das linksventrikulare Myokard gelegt, das ein Partialvolumeneffekt seitens der benachbarten Struktur moglichst gering gehalten wird. Dies erfolgt 13 mal in jeder CT-Schicht. Zusatzlich werden CT-Werte-Anderungen im Lumen des linken Ventrikels wiederum 13 mal bestimmt. Aus der maximalen Steigung der arteriellen myokardialen Zeit-CT-Werte-Kurve wird nach Division durch die Differenz des maximalen CT-Wertes und des Prakontrast-CT-Wertes im linken Ventrikel die myokardiale Perfusion in ml/100 g/min berechnet. Im Rahmen dieser Studie wurden zusatzlich die linksventrikularen Volumina, wie EDV, ESV, SV, bestimmt und daraus AF und HZV errechnet, und zwar aus CT-Aufnahmen des linken Ventrikels, angefertigt in “long axis view”. Bei neun untersuchten Patienten betrugen die gemessenen globalen myokardialen Perfusionswerte 70 ml/100 g/min (minimal 32 und maximal 116 ml/100 g/min). Dieser Mittelwert entspricht 5% des Herz-Zeit-Volumens unter der Annahme, das das mittlere Herzgewicht dieser Patienten 300 g betragt. Die mittels EBCT bestimmte linksventrikulare Muskelmasse betrug in dieser Studie im Durchschnitt 130 g. Der Vergleich der koronarangiographischen Befunde dieser Patienten mit den gemessenen Perfusionswerten zeigte, das es nicht ausreichend ist, nur die absolut gemessenen Perfusionswerte zu betrachten, sondern das diese unter Berucksichtigung des jeweils gemessenen Herz-Zeit-Volumens beurteilt werden mussen. Weitere Studien genau definierter Patienten-gruppen sind erforderlich, um Perfusionsnormwerte unter Ruhebedingungen bei Patienten mit und ohne einer Koronargefaserkrankung zu bestimmen, da vergleichende Auswertungen mit szintigraphischen Verfahren methodisch bedingt problematisch erscheinen. Die bisherigen Ergebnisse zeigen, das es mittels EBCT trotz aufhartungsbedingter Artefakte auch unter klinischen Bedingungen moglich ist, myokardiale Perfusionsmessungen durchzufuhren. Ihre Interpretation mus jedoch vorlaufig unter Berucksichtigung der funktionellen Parameter, einschlieslich der Herzfrequenz, insbesondere jedoch unter Berucksichtigung des Herz-Zeit-Volumens erfolgen.

Classifiers based on parametric or non-parametric learning methods have different advantages and disadvantages. To take advantage of the strengths of both methods, we propose an algorithm that combines a parametric model (logistic regression) with a non-parametric classification method (k-nearest neighbors). This combination is based on a measure of appropriateness that uses a heuristic to decide which of the two components should contribute more to the final classification output. We measure the performance of this combination method on two data sets (one from medical informatics, and one consisting of simulated data) in terms of areas under the ROC curves (AUCs). We are able to demonstrate that our method of combining classifiers exceeds the performance of both individual classifiers taken separately.

Summary Introduction: Understanding the biological processes of signaling pathways as a whole system requires an integrative software environment that has comprehensive capabilities. The environment should include tools for pathway design, visualization, simulation and a knowledge base concerning signaling pathways as one. In this paper we introduce a new integrative environment for the systematic analysis of signaling pathways. Methods: This system includes environments for pathway design, visualization, simulation and a knowledge base that combines biological and modeling information concerning signaling pathways that provides the basic understanding of the biological system, its structure and functioning. The system is designed with a client-server architecture. It contains a pathway designing environment and a simulation environment as upper layers with a relational knowledge base as the underlying layer. Results: The TNFα-mediated NF-κB signal transduction pathway model was designed and tested using our integrative framework. It was also useful to define the structure of the knowledge base. Sensitivity analysis of this specific pathway was performed providing simulation data. Then the model was extended showing promising initial results. Conclusion: The proposed system offers a holistic view of pathways containing biological and modeling data. It will help us to perform biological interpretation of the simulation results and thus contribute to a better understanding of the biological system for drug identification.

Summary Objectives: Marfan syndrome (MFS) is an autosomal dominant inherited connective tissue disorder caused by mutations in the fibrillin-1 (FBN1) gene with variable clinical manifestations in the cardiovascular, musculoskeletal and ocular systems. Methods: Data of molecular genetic analysis and a catalogue of clinical manifestations including aortic elastic parameters were mined in order to (i) assess aortic abnormality before and during medical treatment, and to (ii) identify novel correlations between the genotype and phenotype of the disease using hierarchical cluster analysis and logistic regression analysis. A score measure describing the similarity between a patient’s clinical symptoms and a characteristic phenotype class was introduced. Results: A probabilistic model for monitoring the loss of aortic elasticity was built on merely aortic parameters of 34 patients with classic MFS and 43 control subjects showing a sensitivity of 82% and a specificity of 96%. The clinical phenotypes of 100 individuals with classical or suspected MFS were clustered yielding four different phenotypic expressions. The highest correlation was found between FBN1 missense mutations, which manifested as ectopia lentis, skeletal major and skin minor criteria, and two out of four clustered phenotypes. The probability of the presence of a missense mutation in both phenotype classes is approximately 70%. Conclusions: Monitoring of aortic elastic properties during medical treatment may serve as additional criterion to indicate elective surgical interventions. Genotype-phenotype correlation may contribute to anticipate the clinical consequences of specific FBN1 mutations more comprehensively and may be helpful to identify MFS patients at risk at an early stage of disease.

A core technology in proteomics is mass spectroscopy (MS) that permits the measurement of thousands of proteins/peptides simultanously. Sophisticated data mining methods are necessary to identify highly predictive proteomic biomarker candidates in generated MS spectra that are specific to a certain disease. However, before analysis can be started the preprocessing of raw mass spectra is an essential task, mainly due to the presence of background signals in the spectra such as electrical and chemical noise. In this work we present a new data mining workflow for the identification of proteomic biomarker candidates using mass spectrometry data. The workflow includes two major steps: 1) the preprocessing of raw spectra, and 2) the identification of highly discriminating candidate masses using a 3-step feature selection approach by combining the advantages of efficient filter and effective wrapper techniques. With the proposed workflow we were able to identify putative candidate biomarkers in a lifethreatening human disease using matrix-assisted laser desorption/ionization imaging MS (MALDI-IMS).

High-throughput mass-spectrometry screening has the potential of superior results in detecting early stage cancer than traditional biomarkers. Proteomic data poses novel challenges for data mining, especially concerning the curse of dimensionality. In this paper, we present a 3-step feature selection framework combining the advantages of efficient filter and effective wrapper techniques. We demonstrate the performance of our framework on two SELDITOF- MS data sets for identifying biomarker candidates in ovarian and prostate cancer.

In metabolomics, biomarker discovery is a highly data driven process and requires sophisticated computational methods for the search and prioritization of novel and unforeseen biomarkers in data, typically gathered in preclinical or clinical studies. In particular, the discovery of biomarker candidates from longitudinal cohort studies is crucial for kinetic analysis to better understand complex metabolic processes in the organism during physical activity. In this work we introduce a novel computational strategy that allows to identify and study kinetic changes of putative biomarkers using targeted MS/MS profiling data from time series cohort studies or other cross-over designs. We propose a prioritization model with the objective of classifying biomarker candidates according to their discriminatory ability and couple this discovery step with a novel network-based approach to visualize, review and interpret key metabolites and their dynamic interactions within the network. The application of our method on longitudinal stress test data revealed a panel of metabolic signatures, i.e., lactate, alanine, glycine and the short-chain fatty acids C2 and C3 in trained and physically fit persons during bicycle exercise. We propose a new computational method for the discovery of new signatures in dynamic metabolic profiling data which revealed known and unexpected candidate biomarkers in physical activity. Many of them could be verified and confirmed by literature. Our computational approach is freely available as R package termed BiomarkeR under LGPL via CRAN http://cran.r-project.org/web/packages/BiomarkeR/ .

The search and validation of novel disease biomarkers requires the complementary power of professional study planning and execution, modern profiling technologies and related bioinformatics tools for data analysis and interpretation. Biomarkers have considerable impact on the care of patients and are urgently needed for advancing diagnostics, prognostics and treatment of disease. This survey article highlights emerging bioinformatics methods for biomarker discovery in clinical metabolomics, focusing on the problem of data preprocessing and consolidation, the data-driven search, verification, prioritization and biological interpretation of putative metabolic candidate biomarkers in disease. In particular, data mining tools suitable for the application to omic data gathered from most frequently-used type of experimental designs, such as case-control or longitudinal biomarker cohort studies, are reviewed and case examples of selected discovery steps are delineated in more detail. This review demonstrates that clinical bioinformatics has evolved into an essential element of biomarker discovery, translating new innovations and successes in profiling technologies and bioinformatics to clinical application.

Background The shotgun strategy (liquid chromatography coupled with tandem mass spectrometry) is widely applied for identification of proteins in complex mixtures. This method gives rise to thousands of spectra in a single run, which are interpreted by computational tools. Such tools normally use a protein database from which peptide sequences are extracted for matching with experimentally derived mass spectral data. After the database search, the correctness of obtained peptide-spectrum matches (PSMs) needs to be evaluated also by algorithms, as a manual curation of these huge datasets would be impractical. The target-decoy database strategy is largely used to perform spectrum evaluation. Nonetheless, this method has been applied without considering sensitivity, i.e., only error estimation is taken into account. A recently proposed method termed MUDE treats the target-decoy analysis as an optimization problem, where sensitivity is maximized. This method demonstrates a significant increase in the retrieved number of PSMs for a fixed error rate. However, the MUDE model is constructed in such a way that linear decision boundaries are established to separate correct from incorrect PSMs. Besides, the described heuristic for solving the optimization problem has to be executed many times to achieve a significant augmentation in sensitivity. Results Here, we propose a new method, termed MUMAL, for PSM assessment that is based on machine learning techniques. Our method can establish nonlinear decision boundaries, leading to a higher chance to retrieve more true positives. Furthermore, we need few iterations to achieve high sensitivities, strikingly shortening the running time of the whole process. Experiments show that our method achieves a considerably higher number of PSMs compared with standard tools such as MUDE, PeptideProphet, and typical target-decoy approaches. Conclusion Our approach not only enhances the computational performance, and thus the turn around time of MS-based experiments in proteomics, but also improves the information content with benefits of a higher proteome coverage. This improvement, for instance, increases the chance to identify important drug targets or biomarkers for drug development or molecular diagnostics.

Corpus Callosum Agenesis (CCA), one of the most common congenital anomalies, has uncertain neurodevelopmental outcome, especially when the disease is isolated. To provide parents with informed counselling, it is crucial to identify anatomical markers linked to a predicted outcome early in pregnancy. Quantitative exploration of fetal brains with CCA is rare and has been mostly limited to the study of specific brain structures. Here, we propose a pipeline to analyse fetal brain Magnetic Resonance Imaging (MRI) that is based on diffeomorphic transformation. It consists in two steps: a semi-automatic fetal MRI preprocessing procedure and a pipeline to quantify anatomical deviations from normal development. Following MRI preprocessing, each volumetric fetal brain is compared to an age-matched healthy template brain at a global scale using registration. Deformations are parallel transported to the same space to correct age differences between fetuses. Deformation modes specific to CCA are identified using Principal Component Analysis and classification. The pipeline is tested on retrospectively selected MRIs from 38 healthy fetuses and 73 fetuses with CCA. In accordance with more local analyses, the most relevant deformation mode for classification combines well-known alterations of brains with CCA. This preliminary work is promising for the quantitative exploration of abnormal fetal brains and will be used in the future to identify anatomical features correlated to poor clinical outcome.