Analysis of spectral data in clinical proteomics by use of learning vector quantizers (original) (raw)
Related papers
2008
In the present contribution we propose two recently developed classification algorithms for the analysis of massspectrometric dataçthe supervised neural gas and the fuzzy-labeled self-organizing map. The algorithms are inherently regularizing, which is recommended, for these spectral data because of its high dimensionality and the sparseness for specific problems. The algorithms are both prototype-based such that the principle of characteristic representants is realized. This leads to an easy interpretation of the generated classifcation model. Further, the fuzzy-labeled self-organizing map is able to process uncertainty in data, and classification results can be obtained as fuzzy decisions. Moreover, this fuzzy classification together with the property of topographic mapping offers the possibility of class similarity detection, which can be used for class visualization. We demonstrate the power of both methods for two exemplary examples: the classification of bacteria (listeria types) and neoplastic and non-neoplastic cell populations in breast cancer tissue sections.
Prototype based fuzzy classification in clinical proteomics
International Journal of Approximate Reasoning, 2008
Proteomic profiling based on mass spectrometry is an important tool for studies at the protein and peptide level in medicine and health care. Thereby, the identification of relevant masses, which are characteristic for specific sample states e.g. a disease state is complicated. Further, the classification accuracy and safety is especially important in medicine. The determination of classification models for such high dimensional clinical data is a complex task. Specific methods, which are robust with respect to the large number of dimensions and fit to clinical needs, are required. In this contribution two such methods for the construction of nearest prototype classifiers are compared in the context of clinical proteomic studies, which are specifically suited to deal with such high-dimensional functional data. Both methods are suitable to the adaptation of the underling metric, which is useful in proteomic research to get a problem adequate representation of the clinical data. In addition they allow fuzzy classification and for one of them allows fuzzy classified training data. Both algorithms are investigated in detail with respect to their specific properties. A performance analyzes is taken on real clinical proteomic cancer data in a comparative manner.
PLoS ONE, 2011
The discovery of protein variation is an important strategy in disease diagnosis within the biological sciences. The current benchmark for elucidating information from multiple biological variables is the so called ''omics'' disciplines of the biological sciences. Such variability is uncovered by implementation of multivariable data mining techniques which come under two primary categories, machine learning strategies and statistical based approaches. Typically proteomic studies can produce hundreds or thousands of variables, p, per observation, n, depending on the analytical platform or method employed to generate the data. Many classification methods are limited by an n%p constraint, and as such, require pretreatment to reduce the dimensionality prior to classification. Recently machine learning techniques have gained popularity in the field for their ability to successfully classify unknown samples. One limitation of such methods is the lack of a functional model allowing meaningful interpretation of results in terms of the features used for classification. This is a problem that might be solved using a statistical model-based approach where not only is the importance of the individual protein explicit, they are combined into a readily interpretable classification rule without relying on a black box approach. Here we incorporate statistical dimension reduction techniques Partial Least Squares (PLS) and Principal Components Analysis (PCA) followed by both statistical and machine learning classification methods, and compared them to a popular machine learning technique, Support Vector Machines (SVM). Both PLS and SVM demonstrate strong utility for proteomic classification problems.
Prototype-based fuzzy classification with local relevance for proteomics
Neurocomputing, 2006
In this paper, we extend soft nearest prototype classification by local metric learning and fuzzy classification. Thereby, the metric is determined according to the given classification task. This may be done separately for each prototype or class specific. We apply the method to cancer detection based on proteomic data. r
A major goal of clinical proteomics is the identification of disease-specific biomarker panels that can be obtained from mass spectral analyses of body fluids, such as blood serum or CSF (cerebrospinal fluid). In this paper, we present preliminary results from the application of symbolic machine learning and data mining techniques to three high-throughput proteomic datasets -publicly available ovarian cancer data, pre-release datasets for amyotrophic lateral sclerosis (ALS) and trisomy 21 (Down's syndrome).
Computer Methods and Programs in Biomedicine, 2010
In this study, a pattern recognition system is presented for improving the classification accuracy of MS-spectra by means of gathering information from different MS-spectra intensity regions using a majority vote ensemble combination. The method starts by automatically breaking down all MS-spectra into common intensity regions. Subsequently, the most informative features (m/z values), which might constitute potential significant biomarkers, are extracted from each common intensity region over all the MS-spectra and, finally, normal from ovarian cancer MS-spectra are discriminated using a multi-classifier scheme, with members the Support Vector Machine, the Probabilistic Neural Network and the k-Nearest Neighbour classifiers. Clinical material was obtained from the publicly available ovarian proteomic dataset (8-7-02). To ensure robust and reliable estimates, the proposed pattern recognition system was evaluated using an external cross-validation process. The average overall performance of the system in discriminating normal from cancer ovarian MS-spectra was 97.18% with 98.52% mean sensitivity and 94.84% mean specificity values.
Machine learning methods for predictive proteomics
Briefings in Bioinformatics, 2007
The search for predictive biomarkers of disease from high-throughput mass spectrometry (MS) data requires a complex analysis path. Preprocessing and machine-learning modules are pipelined, starting from raw spectra, to set up a predictive classifier based on a shortlist of candidate features. As a machine-learning problem, proteomic profiling on MS data needs caution like the microarray case. The risk of overfitting and of selection bias effects is pervasive: not only potential features easily outnumber samples by 10 3 times, but it is easy to neglect informationleakage effects during preprocessing from spectra to peaks. The aim of this review is to explain how to build a general purpose design analysis protocol (DAP) for predictive proteomic profiling: we show how to limit leakage due to parameter tuning and how to organize classification and ranking on large numbers of replicate versions of the original data to avoid selection bias. The DAP can be used with alternative components, i.e. with different preprocessing methods (peak clustering or wavelet based), classifiers (e.g. Support Vector Machine) or feature ranking methods (recursive feature elimination or I-Relief). A procedure for assessing stability and predictive value of the resulting biomarkers' list is also provided. The approach is exemplified with experiments on synthetic datasets (from the Cromwell MS simulator) and with publicly available datasets from cancer studies.
Robust SVM-Based Biomarker Selection with Noisy Mass Spectrometric Proteomic Data
Lecture Notes in Computer Science, 2006
Computational analysis of mass spectrometric (MS) proteomic data from sera is of potential relevance for diagnosis, prognosis, choice of therapy, and study of disease activity. To this aim, feature selection techniques based on machine learning can be applied for detecting potential biomarkes and biomaker patterns. A key issue concerns the interpretability and robustness of the output results given by such techniques. In this paper we propose a robust method for feature selection with MS proteomic data. The method consists of the sequentail application of a filter feature selection algorithm, RELIEF, followed by multiple runs of a wrapper feature selection technique based on support vector machines (SVM), where each run is obtained by changing the class label of one support vector. Frequencies of features selected over the runs are used to identify features which are robust with respect to perturbations of the data. This method is tested on a dataset produced by a specific MS technique, called MALDI-TOF MS. Two classes have been artificially generated by spiking. Moreover, the samples have been collected at different storage durations. Leave-one-out cross validation (LOOCV) applied to the resulting dataset, indicates that the proposed feature selection method is capable of identifying highly discriminatory proteomic patterns.
Comparison of Supervised Classification Methods for Protein Profiling in Cancer Diagnosis
Cancer Informatics, 2007
A key challenge in clinical proteomics of cancer is the identifi cation of biomarkers that could allow detection, diagnosis and prognosis of the diseases. Recent advances in mass spectrometry and proteomic instrumentations offer unique chance to rapidly identify these markers. These advances pose considerable challenges, similar to those created by microarray-based investigation, for the discovery of pattern of markers from high-dimensional data, specifi c to each pathologic state (e.g. normal vs cancer). We propose a three-step strategy to select important markers from high-dimensional mass spectrometry data using surface enhanced laser desorption/ionization (SELDI) technology. The fi rst two steps are the selection of the most discriminating biomarkers with a construction of different classifi ers. Finally, we compare and validate their performance and robustness using different supervised classifi cation methods such as Support Vector Machine, Linear Discriminant Analysis, Quadratic Discriminant Analysis, Neural Networks, Classifi cation Trees and Boosting Trees. We show that the proposed method is suitable for analysing high-throughput proteomics data and that the combination of logistic regression and Linear Discriminant Analysis outperform other methods tested.
Data mining techniques for cancer detection using serum proteomic profiling
Artificial Intelligence in Medicine, 2004
Pathological changes in an organ or tissue may be reflected in proteomic patterns in serum. It is possible that unique serum proteomic patterns could be used to discriminate cancer samples from non-cancer ones. Due to the complexity of proteomic profiling, a higher order analysis such as data mining is needed to uncover the differences in complex proteomic patterns. The objectives of this paper are (1) to briefly review the application of data mining techniques in proteomics for cancer detection/diagnosis; (2) to explore a novel analytic method with different feature selection methods; (3) to compare the results obtained on different datasets and that reported by Petricoin et al. in terms of detection performance and selected proteomic patterns. Methods and material: Three serum SELDI MS data sets were used in this research to identify serum proteomic patterns that distinguish the serum of ovarian cancer cases from non-cancer controls. A support vector machine-based method is applied in this study, in which statistical testing and genetic algorithm-based methods are used for feature selection respectively. Leave-one-out cross validation with receiver operating characteristic (ROC) curve is used for evaluation and comparison of cancer detection performance. Results and conclusions: The results showed that (1) data mining techniques can be successfully applied to ovarian cancer detection with a reasonably high performance; (2) the classification using features selected by the genetic algorithm consistently outperformed those selected by statistical testing in terms of accuracy and robustness; (3) the discriminatory features (proteomic patterns) can be very different from one selection method to another. In other words, the pattern selection and its classification efficiency are highly classifier dependent. Therefore, when using data mining techniques, the discrimination of cancer from normal does not depend solely upon the identity and origination of cancer-related proteins.