Artificial neural network classification based on capillary electrophoresis of urinary nucleosides for the clinical diagnosis of tumors

Systematic, Modifying Group-Assisted Strategy Expanding Coverage of Metabolite Annotation in Liquid Chromatography-Mass Spectrometry-Based Nontargeted Metabolomics Studies

From microbes to human beings, nontargeted metabolic profiling by liquid chromatography (LC)-mass spectrometry (MS) has been commonly used to investigate metabolic alterations. Still, a major challenge is the annotation of metabolites from thousands of detected features. The aim of our research was to go beyond coverage of metabolite annotation in common nontargeted metabolomics studies by an integrated multistep strategy applying data-dependent acquisition (DDA)-based ultrahigh-performance liquid chromatography (UHPLC)-high-resolution mass spectrometry (HRMS) analysis followed by comprehensive neutral loss matches for characteristic metabolite modifications and database searches in a successive manner. Using pooled human urine as a model sample for method establishment, we found 22% of the detected compounds having modifying structures. Major types of metabolite modifications in urine were glucuronidation (33%), sulfation (20%), and acetylation (6%). Among the 383 annotated metabolites, 100 were confirmed by standard compounds and 50 modified metabolites not present in common databases such as human metabolite database (HMDB) and Kyoto Encyclopedia of Genes and Genomes (KEGG) were structurally elucidated. Practicability was tested by the investigation of urines from pregnant women diagnosed with gestational diabetes mellitus vs healthy controls. Overall, 83 differential metabolites were annotated and 67% of them were modified metabolites including five previously unreported compounds. To conclude, the systematic modifying group-assisted strategy can be taken as a useful tool to extend the number of annotated metabolites in biological and biomedical nontargeted studies.

Metal oxide-based dispersive solid-phase extraction coupled with mass spectrometry analysis for determination of ribose conjugates in human follicular fluid

Polycystic ovary syndrome (PCOS) is the most common cause of anovulatory infertility. The pathogenesis of PCOS remains unclear and early diagnosis of PCOS is challenging. Follicular fluid provides a unique window in the critical processes during oocyte and follicular maturation, and the metabolic level of follicular fluid has important impact on the developmental potential of oocytes and subsequent embryos. Previous studies demonstrated some modified ribonucleosides in biological fluids were diseases related metabolites. In this respect, analysis of endogenous modified ribonucleosides in follicular fluids will facilitate the investigation of follicular development. Here, we developed a strategy for determination of ribose conjugates from follicular fluid using metal oxide-based dispersive solid-phase extraction (DSPE) coupled with liquid chromatography-multiple reaction monitoring-mass spectrometry analysis (DSPE-LC-MRM-MS/MS). Cerium dioxide (CeO₂) was used to selectively recognize and capture cis-diol containing ribose conjugates from complex biological samples under basic environment. The trapped ribose conjugates were then easily released under acidic environment. The results showed that 50 potential ribose conjugates were detected in follicular fluid by the developed DSPE-LC-MRM-MS/MS method. We then further investigated the contents change of the detected ribose conjugates in follicular fluid from PCOS patients. The results indicated that the follicular fluid from healthy controls and PCOS patients can be clearly differentiated with the partial least squares-discriminate analysis (PLS-DA) based on the detected ribose conjugates. In addition, the contents of 8 ribose conjugates were significantly different between PCOS patients and healthy controls, which could potentially serve as the indicator of PCOS.

Modified nucleosides as biomarkers for early cancer diagnose in exposed populations

There is increasing worldwide interest in developing of markers for tumor diagnosis and identification of individuals who are at high cancer risk. Cancer, like other diseases accompanied by metabolic disorders, causes characteristic effects on cell turnover rate, activity of modifying enzymes, and RNA/DNA modifications. This results in an increased excretion of modified nucleosides in cancer patients. Therefore, for many years modified nucleosides have been suggested as tumor markers. The aim of the study was to elucidate further the usefulness of urinary nucleosides as possible markers at early detection of cancer in persons which are exposed against tumor promoting influences during their working life. Uranium miners are exposed to many kinds of pollutants that can cause health damage even lead to carcinogenesis. We analyzed modified nucleosides in urine samples from 92 miners who are at high risk for lung cancer to assess the levels of nucleosides by a multilayer perceptron (MLP) classifier - a neural network model. Eighteen nucleosides/metabolites were detected with reversed-phase high-pressure liquid chromatography (RP-HPLC). A valid set of urinary metabolites were selected and multivariate statistical technique of multilayer perceptron neural network were applied. In a previous study, MLP shows a sensitivity and specificity of 97 and 85%, respectively. MLP classification including the most relevant markers/nucleosides clearly demonstrates the elevation of RNA metabolism in miners, which is associated with possible malignant disease. We found that there were 30 subjects with early health disorders among 92 uranium workers based on MLP technique using modified nucleosides. The combination of RP-HPLC analysis of modified nucleosides and subsequent MLP analyses represents a promising tool for the development of a non-invasive prediction system and may assist in developing management and surveillance procedures.

Profiling of cis-diol-containing nucleosides and ribosylated metabolites by boronate-affinity organic-silica hybrid monolithic capillary liquid chromatography/mass spectrometry

RNA contains a large number of modified nucleosides. In the metabolic re-exchange of RNA, modified nucleosides cannot be recycled and are thus excreted from cells into biological fluids. Determination of endogenous modified nucleosides in biological fluids may serve as non-invasive cancers diagnostic methods. Here we prepared boronate-affinity organic-silica hybrid capillary monolithic column (BOHCMC) that exhibited excellent selectivity toward the cis-diol-containing compounds. We then used the prepared BOHCMC as the on-line solid-phase microextraction (SPME) column and developed an on-line SPME-LC-MS/MS method to comprehensively profile cis-diol-containing nucleosides and ribosylated metabolites in human urine. Forty-five cis-diol-containing nucleosides and ribosylated metabolites were successfully identified in human urine. And five ribose conjugates, for the first time, were identified existence in human urine in the current study. Furthermore, the relative quantification suggested 4 cis-diol-containing compounds (5′-deoxy-5′-methylthioadensine, N⁴-acetylcytidine, 1-ribosyl-N-propionylhistamine and N²,N²,7-trimethylguanosine) increased more than 1.5 folds in all the 3 types of examined cancers (lung cancer, colorectal cancer, and nasopharyngeal cancer) compared to healthy controls. The on-line SPME-LC-MS/MS method demonstrates a promising method for the comprehensive profiling of cis-diol-containing ribose conjugates in human urines, which provides an efficient strategy for the identification and discovery of biomarkers and may be used for the screening of cancers.

Metabonomic profiling of bladder cancer

Early diagnosis and life-long surveillance are clinically important to improve the long-term survival of bladder cancer patients. Currently, a noninvasive biomarker that is as sensitive and specific as cystoscopy in detecting bladder tumors is lacking. Metabonomics is a complementary approach for identifying perturbed metabolic pathways in bladder cancer. Significant progress has been made using modern metabonomic techniques to characterize and distinguish bladder cancer patients from control subjects, identify marker metabolites, and shed insights on the disease biology and potential therapeutic targets. With its rapid development, metabonomics has the potential to impact the clinical management of bladder cancer patients in the future by revolutionizing the diagnosis and life-long surveillance strategies and stratifying patients for diagnostic, surgical, and therapeutic clinical trials. An introduction to metabonomics, typical metabonomic workflow, and critical evaluation of metabonomic investigations in identifying biomarkers for the diagnosis of bladder cancer are presented.

Nontargeted modification-specific metabolomics study based on liquid chromatography-high-resolution mass spectrometry

Modifications of genes and proteins have been extensively studied in systems biology using comprehensive analytical strategies. Although metabolites are frequently modified, these modifications have not been studied using -omics approaches. Here a general strategy for the nontargeted profiling of modified metabolites, which we call "nontargeted modification-specific metabolomics", is reported. A key aspect of this strategy was the combination of in-source collision-induced dissociation liquid chromatography-mass spectrometry (LC-MS) and global nontargeted LC-MS-based metabolomics. Characteristic neutral loss fragments that are specific for acetylation, sulfation, glucuronidation, glucosidation, or ribose conjugation were reproducibly detected using human urine as a model specimen for method development. The practical application of this method was demonstrated by profiling urine samples from liver cirrhosis patients. Approximately 900 features were identified as modified endogenous metabolites and xenobiotics. Moreover, this strategy supports the identification of compounds not included in traditional metabolomics databases (HMDB, Metlin, and KEGG), which are currently referred to as "unknowns" in metabolomics projects. Nontargeted modification-specific metabolomics opens a new perspective in systems biology.

Neural network and SVM classifiers accurately predict lipid binding proteins, irrespective of sequence homology

Due to the central roles of lipid binding proteins (LBPs) in many biological processes, sequence based identification of LBPs is of great interest. The major challenge is that LBPs are diverse in sequence, structure, and function which results in low accuracy of sequence homology based methods. Therefore, there is a need for developing alternative functional prediction methods irrespective of sequence similarity. To identify LBPs from non-LBPs, the performances of support vector machine (SVM) and neural network were compared in this study. Comprehensive protein features and various techniques were employed to create datasets. Five-fold cross-validation (CV) and independent evaluation (IE) tests were used to assess the validity of the two methods. The results indicated that SVM outperforms neural network. SVM achieved 89.28% (CV) and 89.55% (IE) overall accuracy in identification of LBPs from non-LBPs and 92.06% (CV) and 92.90% (IE) (in average) for classification of different LBPs classes. Increasing the number and the range of extracted protein features as well as optimization of the SVM parameters significantly increased the efficiency of LBPs class prediction in comparison to the only previous report in this field. Altogether, the results showed that the SVM algorithm can be run on broad, computationally calculated protein features and offers a promising tool in detection of LBPs classes. The proposed approach has the potential to integrate and improve the common sequence alignment based methods.

Understanding the undelaying mechanism of HA-subtyping in the level of physic-chemical characteristics of protein

The evolution of the influenza A virus to increase its host range is a major concern worldwide. Molecular mechanisms of increasing host range are largely unknown. Influenza surface proteins play determining roles in reorganization of host-sialic acid receptors and host range. In an attempt to uncover the physic-chemical attributes which govern HA subtyping, we performed a large scale functional analysis of over 7000 sequences of 16 different HA subtypes. Large number (896) of physic-chemical protein characteristics were calculated for each HA sequence. Then, 10 different attribute weighting algorithms were used to find the key characteristics distinguishing HA subtypes. Furthermore, to discover machine leaning models which can predict HA subtypes, various Decision Tree, Support Vector Machine, Naïve Bayes, and Neural Network models were trained on calculated protein characteristics dataset as well as 10 trimmed datasets generated by attribute weighting algorithms. The prediction accuracies of the machine learning methods were evaluated by 10-fold cross validation. The results highlighted the frequency of Gln (selected by 80% of attribute weighting algorithms), percentage/frequency of Tyr, percentage of Cys, and frequencies of Try and Glu (selected by 70% of attribute weighting algorithms) as the key features that are associated with HA subtyping. Random Forest tree induction algorithm and RBF kernel function of SVM (scaled by grid search) showed high accuracy of 98% in clustering and predicting HA subtypes based on protein attributes. Decision tree models were successful in monitoring the short mutation/reassortment paths by which influenza virus can gain the key protein structure of another HA subtype and increase its host range in a short period of time with less energy consumption. Extracting and mining a large number of amino acid attributes of HA subtypes of influenza A virus through supervised algorithms represent a new avenue for understanding and predicting possible future structure of influenza pandemics.

Quantitative analysis of adenosine using liquid chromatography/atmospheric pressure chemical ionization-tandem mass spectrometry (LC/APCI-MS/MS)

Adenosine-secreting cellular brain implants constitute a promising therapeutic approach for the treatment of epilepsy. To engineer neural stem cells for therapeutic adenosine delivery, a reliable and fast analytical method is necessary to quantify cell-based adenosine release. Here we describe the development, optimization and validation of adenosine measurement using liquid chromatography-atmospheric pressure chemical ionization-tandem mass spectrometry (LC-APCI-MS/MS). LC-MS/MS in positive ion mode used selected reaction monitoring at m/z of 268.2/136.1 and 302.2/170.0 for adenosine and the internal standard, respectively. The bias was within 15% of the nominal value and evaluation of precision showed a relative standard deviation lower than 15% for all measured concentrations. The lower limit of quantification of adenosine was 15.6 ng/ml. Freeze and thaw stability and processed sample stability also fulfilled the acceptance criteria. Evaluation of the matrix effect showed that the method is not affected by relative matrix effects. The major advantages of this method are the absence of an extraction phase and the combination of the high selectivity and sensitivity characteristic for the LC-MS/MS technique, with a short run time of 4.5 min. These results demonstrate that this method is a useful tool to measure adenosine concentrations in culture medium released from stem cells in vitro.

Application of artificial neural networks in the prediction of product distribution in electrophoretically mediated microanalysis

The successful application of artificial neural networks toward the prediction of product distribution in electrophoretically mediated microanalysis is presented. To illustrate this concept, we examined the factors and levels required for optimization of reaction conditions for the conversion of nicotinamide adenine dinucleotide to nicotinamide adenine dinucleotide, reduced form by glucose-6-phosphate dehydrogenase in the conversion of glucose-6-phosphate to 6-phosphogluconate. A full factorial experimental design examining the factors voltage, enzyme concentration, and mixing time of reaction was utilized as input-output data sources for suitable artificial neural networks training for prediction purposes. This approach proved successful in predicting optimal values in a reduced number of experiments. Model validation addressing the extent of reaction and product ratios were subsequently determined experimentally in replicate analyses, with results shown to be in good agreement (<10% discrepancy difference) with predicted data.

Urinary nucleosides based potential biomarker selection by support vector machine for bladder cancer recognition

BACKGROUND

Urinary nucleosides are potential biomarkers for many kinds of cancers. But up to now, it has been little focused in bladder cancer recognition. The aim of present study is try to validate the potential of urinary nucleoside as biomarker for bladder cancer diagnosis by finding out some urinary nucleosides with good discriminative performance for bladder cancer recognition in urinary nucleoside profile.

METHODS

20 urinary samples for cancer and the same number for control are collected and treated by capillary electrophoresis-mass spectrometry experiments to achieve urinary nucleoside profile, in which 44 peaks were integrated and the ratios of the relative peak area to the concentration of urinary creatinine were used as features to describe all samples. Support vector machine based recursive feature elimination (SVM-RFE) and a new feature selection method called support vector machine based partial exhaustive search algorithm (SVM-PESA) were used for biomarker identification and seeking optimal feature subsets for bladder cancer recognition.

RESULTS

Based on the urinary nucleoside profile, 22 optimal feature subsets consist of 3-4 features were found with 95% 5-fold cross validation accuracy, 100% sensitivity and 90% specificity by SVM-PESA, whose performance were much better than that of optimal feature subset selected by SVM-RFE. By analyzing the statistical histogram of features' appearance frequency in several best feature subsets, urinary nucleosides with m/z 317, 290 and 304 were thought as potential biomarkers for bladder cancer recognition.

CONCLUSIONS

These results indicated urinary nucleosides may be useful as tumor biomarkers for bladder cancer, and the new method for biomarker selection is effective.

Multilayer perceptron tumour diagnosis based on chromatography analysis of urinary nucleosides

Nucleosides in human urine are of interest as a biochemical marker for cancer, acquired immune deficiency syndrome (AIDS) and the whole-body turnover of RNAs. A reversed-phase high-performance liquid chromatography (RP-HPLC) method with photodiode-array detection was used to quantitatively analyze urinary normal and modified nucleosides. 55 persons with malignant tumors of various types, 13 persons with benign tumors and 41 healthy controls were investigated within a clinical intervention study. Artificial neural networks (ANN) have been used as a practical pattern recognition tool to distinguish cancer patients from healthy persons. Using a multilayer perceptron (MPL), a specificity of 85%, and a sensitivity of 97% in differentiation between tumor patients and healthy persons was achieved. The differentiation between benign and malignant tumors had a sensitivity of 60% and a specificity of 84%. These results verify the usefulness of ANN and the RP-HPLC method for tumor recognition in agreement with existing studies.

Novel approach for developing urinary nucleosides profile by capillary electrophoresis–mass spectrometry

A simple, rapid and efficient capillary electrophoresis-mass spectrometry (CE-MS) method was developed to analyze urinary nucleosides for the first time. The composition of CE buffer and MS parameters were systematically optimized. The optimum buffer was 150 mM acetic acid containing 15% methanol and 15% ethanol. The optimum MS parameters were: methanol containing 0.5% acetic acid was selected as the sheath liquid and the flow rate was 5 microL/min; the flow rate and temperature of drying gas were 6L/min and 150 degrees C, respectively; the pressure of nebulizing gas was 2 psig; and the fragmentor and ESI voltage were 100 V and 4000 V, respectively. Under the optimum CE-MS conditions, the urinary nucleosides were separated within 18 min. The linearity between the relative peak areas and the corresponding concentration of nine nucleosides markers were excellent. The limits of detection (S/N=3) of markers were 0.00862-3.82 nmol/mL. The optimum CE-MS method was applied to analyze urine from 20 bladder cancer patients and 20 healthy volunteers. Considering the standards of many nucleosides cannot be obtained, it is not the ratios of the concentrations of nucleosides to that of creatinine in the literatures, but the ratios of the relative peak area of nucleosides to the concentration of creatinine that used for pattern recognition. And, the statistical analysis result indicated this method was feasible.

Modeling force-velocity relation in skeletal muscle isotonic contraction using an artificial neural network

The aim of this study is to design an artificial neural network (ANN) to model force-velocity relation in skeletal muscle isotonic contraction. We obtained the data set, including physiological and morphometric parameters, by myography and morphometric measurements on frog gastrocnemius muscle. Then, we designed a multilayer perceptron ANN, the inputs of which are muscle volume, muscle optimum length, tendon length, preload, and afterload. The output of the ANN is contraction velocity. The experimental data were divided randomly into two parts. The first part was used to train the ANN. In order to validate the model, the second part of experimental data, which was not used in training, was employed to the ANN and then, its output was compared with Hill model and the experimental data. The behavior of ANN in high forces was more similar to experimental data, but in low forces the Hill model had better results. Furthermore, extrapolation of ANN performance showed that our model is more or less able to simulate eccentric contraction. Our results indicate that ANNs represent a powerful tool to capture some essential features of muscle isotonic contraction.

Modified nucleosides: an accurate tumour marker for clinical diagnosis of cancer, early detection and therapy control

Modified nucleosides, regarded as indicators for the whole-body turnover of RNAs, are excreted in abnormal amounts in the urine of patients with malignancies. To test their usefulness as tumour markers and to compare them with the conventional tumour markers, fractionated urine samples were analysed using chromatography. The excretion patterns of nucleosides of 68 cancer patients with malignant and benign tumours and 41 healthy controls have been studied. Significant elevations in the total sum and the concentrations of at least three (or four) of indicator nucleosides cytidine, pseudouridine, 2-pyridone-5-carboxamide-N1-ribofuranoside, N2,N2-dimethylguanine, 1-methylguanosine, 2-methylguanosine and 1-methyladenosine indicate a tumour with a sensitivity of 54% (77%) and a specificity of 86% (98%). Using an artificial neural network analysis, a sensitivity of 97% and a specificity of 85% were achieved in differentiating between tumour and control volunteers. The comparison with carcinoembryonic antigen, cancer antigen 15-3 und tissue polypeptide antigen indicates that urinary nucleosides may be useful tumour markers. This study suggests that the simultaneous determination of modified nucleosides and creatinine in urine samples of patients with cancer leads to an advantage to current methods and is a useful method to detect cancer early and to control the success of therapy.

Urinary nucleosides as biological markers for patients with colorectal cancer

AIM: Fourteen urinary nucleosides, primary degradation products of tRNA, were evaluated to know the potential as biological markers for patients with colorectal cancer.

METHODS: The concentrations of 14 kinds of urinary nucleosides from 52 patients with colorectal cancer, 10 patients with intestinal villous adenoma and 60 healthy adults were determined by column switching high performance liquid chromatography method.

RESULTS: The mean levels of 12 kinds of urinary nucleosides (except uridine and guanosine) in the patients with colorectal cancer were significantly higher than those in patients with intestinal villous adenoma or the healthy adults. Using the levels of 14 kinds of urinary nucleosides as the data vectors for principal component analysis, 71% (37/52) patients with colorectal cancer were correctly classified from healthy adults, in which the identification rate was much higher than that of CEA method (29%). Only 10% (1/10) of patients with intestinal villous adenoma were indistinguishable from patients with colorectal cancer. The levels of m1G, Pseu and m1A were positively related with tumor size and Duke’s stages of colorectal cancer. When monitoring the changes in urinary nucleoside concentrations of patients with colorectal cancer associated with surgery, it was found that the overall correlations with clinical assessment were 84% (27/32) and 91% (10/11) in response group and progressive group, respectively.

CONCLUSION: These findings indicate that urinary nucleosides determined by column switching high performance liquid chromatography method may be useful as biological markers for colorectal cancer.

Micellar electrokinetic chromatographic and capillary zone electrophoretic methods for screening urinary biomarkers of human disorders: a critical review of the state-of-the-art

Human urine plays a central role in clinical diagnostic being one of the most-frequently used body fluid for detection of biological markers. Samples from patients with different diseases display patterns of biomarkers that differ significantly from those obtained from healthy subjects. The availability of fast, reproducible, and easy-to-apply analytical techniques that would allow identification of a large number of these analytes is thus highly desiderable since they may provide detailed information about the progression of a pathological process. From among the variety of methods so far applied for the determination of urinary metabolites, capillary electrophoresis, both in the capillary zone electrophoresis (CZE) and micellar electrokinetic chromatography (MEKC) modes, represents a robust and reliable analytical tool widely used in this area. The aim of the present article is to focus the interest of the reader on recent applications of MEKC and CZE in the field of urinary biomarkers and to discuss advantages and/or limitations of each mode.

Clinical significance and prognostic value of urinary nucleosides in breast cancer patients

OBJECTIVE

Thirteen urinary nucleosides, primarily degradation products of tRNA, were evaluated as potential tumor markers for breast cancer patients.

DESIGN AND METHODS

The micellar electrokinetic chromatography (MEKC) method has been used to analyze the urinary nucleosides in 41 healthy controls, 20 patients with benign breast tumors, and 26 breast cancer patients.

RESULTS

Urinary nucleoside concentrations of breast cancer patients were found to increase significantly compared to those of patients with benign breast tumors and healthy controls. By using 13 nucleoside concentrations as data vectors for principal component analysis (PCA), 73% (19/26) of breast cancer patients were correctly identified from healthy controls, while only 20% (4/20) of patients with benign breast tumors were indistinguishable from breast cancer patients. The mean level of all forms of urinary nucleosides in patients with metastatic breast cancer was higher than that in patients with primary breast cancer. The levels of modified nucleosides tended to decrease and return to normal after surgery.

CONCLUSION

The results indicate that urinary nucleosides may be useful as tumor markers for breast cancer.

Analysis of urinary nucleosides. IV. Identification of urinary purine nucleosides by liquid chromatography/electrospray mass spectrometry

Modified urinary nucleosides are potentially invaluable in cancer diagnosis. High-performance liquid chromatography (HPLC) was combined with full scan mass spectrometry (MS), tandem mass spectrometry and MSn analysis in order to identify purine nucleosides purified from urine. UV peaks evident in the chromatogram were examined by the various mass spectrometric techniques and adenosine, 1-methyladenosine, xanthosine, N1-methylguanosine, N2-methylguanosine, N2,N2-dimethylguanosine, N2,N2,N7-trimethylguanosine, inosine, and 1-methylinosine were each identified in the urine samples from cancer patients. The benefits of the use of LC/MS compared with HPLC alone are discussed.

Urinary nucleosides as potential tumor markers evaluated by learning vector quantization

Modified nucleosides were recently presented as potential tumor markers for breast cancer. The patterns of the levels of urinary nucleosides are different for tumor bearing individuals and for healthy individuals. Thus, a powerful pattern recognition method is needed. Although backpropagation (BP) neural networks are becoming increasingly common in medical literature for pattern recognition, it has been shown that often-superior methods exist like learning vector quantization (LVQ) and support vector machines (SVM). The aim of this feasibility study is to get an indication of the performance of urinary nucleoside levels evaluated by LVQ in contrast to the evaluation the popular BP and SVM networks. Urine samples were collected from female breast cancer patients and from healthy females. Twelve different ribonucleosides were isolated and quantified by a high performance liquid chromatography (HPLC) procedure. LVQ, SVM and BP networks were trained and the performance was evaluated by the classification of the test sets into the categories "cancer" and "healthy". All methods showed a good classification with a sensitivity ranging from 58.8 to 70.6% at a specificity of 88.4-94.2% for the test patterns. Although the classification performance of all methods is comparable, the LVQ implementations are superior in terms of more qualitative features: the results of LVQ networks are more reproducible, as the initialization is deterministic. The LVQ networks can be trained by unbalanced sizes of the different classes. LVQ networks are fast during training, need only few parameters adjusted for training and can be retrained by patterns of "local individuals". As at least some of these features play an important role in an implementation into a medical decision support system, it is recommended to use LVQ for an extended study.

Artificial neural network classification based on high-performance liquid chromatography of urinary and serum nucleosides for the clinical diagnosis of cancer

Nucleosides in human urine and serum have frequently been studied as a possible biomedical marker for cancer, acquired immune deficiency syndrome (AIDS) and the whole-body turnover of RNAs. Fifteen normal and modified nucleosides were determined in 69 urine and 42 serum samples using high-performance liquid chromatography (HPLC). Artificial neural networks have been used as a powerful pattern recognition tool to distinguish cancer patients from healthy persons. The recognition rate for the training set reached 100%. In the validating set, 95.8 and 92.9% of people were correctly classified into cancer patients and healthy persons when urine and serum were used as the sample for measuring the nucleosides. The results show that the artificial neural network technique is better than principal component analysis for the classification of healthy persons and cancer patients based on nucleoside data.

Quantitation of intracellular concentration of a delivered morpholino oligomer by capillary electrophoresis-laser- induced fluorescence: correlation with upregulation of luciferase gene expression

Antisense oligonucleotides have shown great promise over the past several years as viable drugs to combat various forms of cancer and viral diseases. However, quantitative detection to monitor cellular association is difficult using conventional methods such as radiolabeling of the oligonucleotide or fluorescence confocal microscopy. In this paper quantitation of intracellular concentration of the morpholino oligonucleotide is investigated using capillary electrophoresis coupled with laser-induced fluorescence detection (CE-LIF). HeLa cells, which produce luciferase as the antisense oligomer enters the cell, were scrape-loaded with varying concentrations of the morpholino antisense. The intracellular antisense concentration measured by CE-LIF was found to correlate with those obtained with the cellular functional assay based on upregulation of luciferase. Intracellular concentrations of the antisense were found to be in the range of 6 to 29 nmol/g total cell protein, depending on the amounts that were scrape-loaded. To our best knowledge, this is the first reported quantitative correlation between delivered antisense concentration in a cell extract and the subsequent antisense upregulation of gene expression.

Capillary electrophoretic profiling and pattern recognition analysis of urinary nucleosides from uterine myoma and cervical cancer patients

Capillary electrophoretic (CE) profiling analysis combined with pattern recognition methods is described for the correlation between urinary nucleoside profiles and uterine cervical cancer. Nucleosides were extracted from urine specimens by solid-phase extraction in affinity mode using phenylboronic acid gel. CE separation was carried out with an uncoated fused-silica capillary (570 mm x 50 microm I.D.) maintained at 20 degrees C, using 25 mM borate-42.5 mM phosphate buffer (pH 6.7) containing 200 mM sodium dodecyl sulfate as the run buffer under the applied voltage of 20 kV. A total of 15 nucleosides were positively identified in urine samples (2 ml) from eight uterine myoma (benign tumor group), 10 uterine cervical cancer (malignant tumor group) patients and 10 healthy females (normal group) studied. The star symbol plots drawn based on each mean concentration of nucleosides normalized to that in normal group enabled one to discriminate malignant and benign groups from normal group. In addition, canonical discriminant analysis performed on the nucleoside data of 28 individual urine specimens correctly classified into three separate clusters according to groups in the canonical plot.

Excretion pattern investigation of urinary normal and modified nucleosides of breast cancer patients by RP-HPLC and factor analysis method

Modified nucleosides, formed post-transcriptionally in RNA by a number of modification enzymes, are excreted in abnormal levels in the urine of patients with malignant tumors. To test their usefulness as tumor markers, and to compare them with the conventional tumor markers, a reversed-phase high-performance liquid chromatographic (RP-HPLC) method and a factor analysis method have been used to study the excretion pattern of nucleosides of breast cancer patients. A clear cut differentiation of the breast cancer group and the healthy individuals in two clusters without overlapping was obtained.

Application of capillary electrophoresis in clinical chemistry: the clinical value of urinary modified nucleosides

Urinary modified nucleosides were determined by capillary electrophoresis using a 300 mM SDS-25 mM sodium tetraborate-50 mM sodium dihydrogenphosphate buffer. The nucleosides were extracted from urine by phenylboronate affinity gel chromatography. In cancer patients the levels of the modified nucleosides are generally elevated. By an artificial neural network method breast cancer patients were differentiated from normal individuals, which indicates that the modified nucleosides could be of clinical value as tumor markers.

Sniffing out the truth: clinical diagnosis using the electronic nose

Recently the use of smell in clinical diagnosis has been rediscovered due to major advances in odour sensing technology and artificial intelligence (AI). It was well known in the past that a number of infectious or metabolic diseases could liberate specific odours characteristic of the disease stage. Later chromatographic techniques identified an enormous number of volatiles in human clinical specimens that might serve as potential disease markers. "Artificial nose" technology has been employed in several areas of medical diagnosis, including rapid detection of tuberculosis (TB), Helicobacter pylori (HP) and urinary tract infections (UTI). Preliminary results have demonstrated the possibility of identifying and characterising microbial pathogens in clinical specimens. A hybrid intelligent model of four interdependent "tools", odour generation "kits", rapid volatile delivery and recovery systems, consistent low drift sensor performance and a hybrid intelligent system of parallel neural networks (NN) and expert systems, have been applied in gastric, pulmonary and urine diagnosis. Initial clinical tests have shown that it may be possible in the near future to use electronic nose technology not only for the rapid detection of diseases such as peptic ulceration, UTI, and TB but also for the continuous dynamic monitoring of disease stages. Major advances in information and gas sensor technology could enhance the diagnostic power of future bio-electronic noses and facilitate global surveillance models of disease control and management.