Classification of individual lung cancer cell lines based on DNA methylation markers: use of linear discriminant analysis and artificial neural networks

Characterizing human mesenchymal stromal cells' immune-modulatory potency using targeted lipidomic profiling of sphingolipids

Cell therapies are expected to increase over the next decade owing to increasing demand for clinical applications. Mesenchymal stromal cells (MSCs) have been explored to treat a number of diseases, with some successes in early clinical trials. Despite early successes, poor MSC characterization results in lessened therapeutic capacity once in vivo. Here, we characterized MSCs derived from bone marrow (BM), adipose tissue and umbilical cord tissue for sphingolipids (SLs), a class of bioactive lipids, using liquid chromatography/tandem mass spectrometry. We found that ceramide levels differed based on the donor's sex in BM-MSCs. We detected fatty acyl chain variants in MSCs from all three sources. Linear discriminant analysis revealed that MSCs separated based on tissue source. Principal component analysis showed that interferon-γ-primed and unstimulated MSCs separated according to their SL signature. Lastly, we detected higher ceramide levels in low indoleamine 2,3-dioxygenase MSCs, indicating that sphingomyelinase or ceramidase enzymatic activity may be involved in their immune potency.

Application and performance of artificial intelligence technology in cytopathology

Deep learning algorithms and artificial intelligence (AI) are making great progress in their capacity to evaluate and interpret image data recent advancements in computer vision and machine learning. The first use of AI in a pathology lab was in cytopathology, when a computer-assisted Pap test screening was created. Initially designed to diagnose rather than screen, there was a lot of disagreement concerning their wide use to clinical specimens. However, whole-slide imaging of both gynaecological and non-gynaecological histopathology have been the subject of recent AI work. An overview of the literature on AI in cytopathology is provided in this brief review. To be more precise, it intends to emphasize the relevance of applications of AI algorithms to gynaecological and non-gynaecologic cytology. Between January 2000 and December 2021, a search on artificial intelligence in cytopathology was conducted in several well-known databases, including PubMed, Web of Science, Scopus, Embase, and Google Scholar. Only full-text papers that could be accessed online were evaluated.

Neural Network Aided Detection of Huntington Disease

Huntington Disease (HD) is a degenerative neurological disease that causes a significant impact on the quality of life of the patient and eventually death. In this paper we present an approach to create a biomarker using as an input DNA CpG methylation data to identify HD patients. DNA CpG methylation is a well-known epigenetic marker for disease state. Technological advances have made it possible to quickly analyze hundreds of thousands of CpGs. This large amount of information might introduce noise as potentially not all DNA CpG methylation levels will be related to the presence of the illness. In this paper, we were able to reduce the number of CpGs considered from hundreds of thousands to 237 using a non-linear approach. It will be shown that using only these 237 CpGs and non-linear techniques such as artificial neural networks makes it possible to accurately differentiate between control and HD patients. An underlying assumption in this paper is that there are no indications suggesting that the process is linear and therefore non-linear techniques, such as artificial neural networks, are a valid tool to analyze this complex disease. The proposed approach is able to accurately distinguish between control and HD patients using DNA CpG methylation data as an input and non-linear forecasting techniques. It should be noted that the dataset analyzed is relatively small. However, the results seem relatively consistent and the analysis can be repeated with larger data-sets as they become available.

Computational methods and next-generation sequencing approaches to analyze epigenetics data: Profiling of methods and applications

Epigenetics is mainly comprised of features that regulate genomic interactions thereby playing a crucial role in a vast array of biological processes. Epigenetic mechanisms such as DNA methylation and histone modifications influence gene expression by modulating the packaging of DNA in the nucleus. A plethora of studies have emphasized the importance of analyzing epigenetics data through genome-wide studies and high-throughput approaches, thereby providing key insights towards epigenetics-based diseases such as cancer. Recent advancements have been made towards translating epigenetics research into a high throughput approach such as genome-scale profiling. Amongst all, bioinformatics plays a pivotal role in achieving epigenetics-related computational studies. Despite significant advancements towards epigenomic profiling, it is challenging to understand how various epigenetic modifications such as chromatin modifications and DNA methylation regulate gene expression. Next-generation sequencing (NGS) provides accurate and parallel sequencing thereby allowing researchers to comprehend epigenomic profiling. In this review, we summarize different computational methods such as machine learning and other bioinformatics tools, publicly available databases and resources to identify key modifications associated with epigenetic machinery. Additionally, the review also focuses on understanding recent methodologies related to epigenome profiling using NGS methods ranging from library preparation, different sequencing platforms and analytical techniques to evaluate various epigenetic modifications such as DNA methylation and histone modifications. We also provide detailed information on bioinformatics tools and computational strategies responsible for analyzing large scale data in epigenetics.

Radial Basis Function Artificial Neural Network for the Investigation of Thyroid Cytological Lesions

Objective

This study investigates the potential of an artificial intelligence (AI) methodology, the radial basis function (RBF) artificial neural network (ANN), in the evaluation of thyroid lesions. Study Design. The study was performed on 447 patients who had both cytological and histological evaluation in agreement. Cytological specimens were prepared using liquid-based cytology, and the histological result was based on subsequent surgical samples. Each specimen was digitized; on these images, nuclear morphology features were measured by the use of an image analysis system. The extracted measurements (41,324 nuclei) were separated into two sets: the training set that was used to create the RBF ANN and the test set that was used to evaluate the RBF performance. The system aimed to predict the histological status as benign or malignant.

Results

The RBF ANN obtained in the training set has sensitivity 82.5%, specificity 94.6%, and overall accuracy 90.3%, while in the test set, these indices were 81.4%, 90.0%, and 86.9%, respectively. Algorithm was used to classify patients on the basis of the RBF ANN, the overall sensitivity was 95.0%, the specificity was 95.5%, and no statistically significant difference was observed.

Conclusion

AI techniques and especially ANNs, only in the recent years, have been studied extensively. The proposed approach is promising to avoid misdiagnoses and assists the everyday practice of the cytopathology. The major drawback in this approach is the automation of a procedure to accurately detect and measure cell nuclei from the digitized images.

Prognosis in pathology: Are we "prognosticating" or only establishing correlations between independent variables and survival? A study with various analytics cautions about the overinterpretation of statistical results

Survival data from 225 patients with resected pulmonary typical carcinoids were analyzed with Kaplan-Meier statistics (K-M) and "deep learning" methods to illustrate the difference between establishing "correlations" and "prognostications". Cases were stratified into G1 and G2 classes using a ≤5% Ki-67% cut-point. Overall survival, number of patients at risk and 95% confidence intervals (CI) were estimated for the two classes. Seven neural network models (NN) were developed with GMDH Shell 3.8.2 and Statgraphics Centurion 18.1 software, using variable prior probabilities and different numbers of training vs testing cases. The NNs used age, sex, and pTNM, G1 and G2 as input neurons and "alive" and "dead" as output neurons. Areas under the curve (AUC) and other performance measures were evaluated for all models. Log-rank test showed a significant difference in overall survival between G1 and G2 (p < 0.001). However, 95% CI estimates showed considerable variability in survival at different time intervals. Including the number of patients at risk at different time intervals showed that most G2 patients had been censored by 100 weeks. The NN models provided variable "prognostications", with AUC ranging from 0.5 to 1 and variability in the sensitivity, specificity, and other performance measures. The results illustrate the limitations of survival statistics and NNs in predicting the prognosis of individual patients. The need for pathologists not to overinterpret the finding of significant correlations as "prognostic" or "predictive" for individual patients is discussed.

The Journal of Molecular Diagnostics: 20 Years of Clinical Innovation

This guest editorial highlights 20 years of clinical innovation in JMD.

Parallel gene selection and dynamic ensemble pruning based on Affinity Propagation

Gene selection and sample classification based on gene expression data are important research areas in bioinformatics. Selecting important genes closely related to classification is a challenging task due to high dimensionality and small sample size of microarray data. Extended rough set based on neighborhood has been successfully applied to gene selection, as it can select attributes without redundancy and deal with numerical attributes directly. However, the computation of approximations in rough set is extremely time consuming. In this paper, in order to accelerate the process of gene selection, a parallel computation method is proposed to calculate approximations of intersection neighborhood rough set. Furthermore, a novel dynamic ensemble pruning approach based on Affinity Propagation clustering and dynamic pruning framework is proposed to reduce memory usage and computational cost. Experimental results on three Arabidopsis thaliana biotic and abiotic stress response datasets demonstrate that the proposed method can obtain better classification performance than ensemble method with gene pre-selection.

HPV Detection Methods

Human papilloma viruses (HPVs) have been acknowledged to be the leading risk factor of cervical intra-epithelial lesion creation (CIN) and cervical cancer development (CxCa). Many different techniques have been created and utilized in HPV detection and monitoring with a vast amount of them being commercialized and few of them integrated in official screening strategies. A growing trend for DNA typing of the 14 most commonly accepted high risk HPV types has been introduced, supporting that in many cases molecular testing could replace classic morphologic diagnostic routines, even though DNA detection has lower specificity than other molecular and morphology tests. However, there have been limited attempts in combining data from all different techniques to provide efficient patient triaging schemes, since, apart from the obvious increase of patient cost, the amount of data and its interpretation in patient management has been impossible. Complex computer based clinical support decision systems, many of which are based on artificial intelligence may abolish these limitations.

Application of artificial neural network model combined with four biomarkers in auxiliary diagnosis of lung cancer

The purpose of the study was to explore the application of artificial neural network model in the auxiliary diagnosis of lung cancer and compare the effects of back-propagation (BP) neural network with Fisher discrimination model for lung cancer screening by the combined detections of four biomarkers of p16, RASSF1A and FHIT gene promoter methylation levels and the relative telomere length. Real-time quantitative methylation-specific PCR was used to detect the levels of three-gene promoter methylation, and real-time PCR method was applied to determine the relative telomere length. BP neural network and Fisher discrimination analysis were used to establish the discrimination diagnosis model. The levels of three-gene promoter methylation in patients with lung cancer were significantly higher than those of the normal controls. The values of Z(P) in two groups were 2.641 (0.008), 2.075 (0.038) and 3.044 (0.002), respectively. The relative telomere lengths of patients with lung cancer (0.93 ± 0.32) were significantly lower than those of the normal controls (1.16 ± 0.57), t = 4.072, P < 0.001. The areas under the ROC curve (AUC) and 95 % CI of prediction set from Fisher discrimination analysis and BP neural network were 0.670 (0.569-0.761) and 0.760 (0.664-0.840). The AUC of BP neural network was higher than that of Fisher discrimination analysis, and Z(P) was 0.76. Four biomarkers are associated with lung cancer. BP neural network model for the prediction of lung cancer is better than Fisher discrimination analysis, and it can provide an excellent and intelligent diagnosis tool for lung cancer.

Artificial Neural Networks as Decision Support Tools in Cytopathology: Past, Present, and Future

OBJECTIVE

This study aims to analyze the role of artificial neural networks (ANNs) in cytopathology. More specifically, it aims to highlight the importance of employing ANNs in existing and future applications and in identifying unexplored or poorly explored research topics.

STUDY DESIGN

A systematic search was conducted in scientific databases for articles related to cytopathology and ANNs with respect to anatomical places of the human body where cytopathology is performed. For each anatomic system/organ, the major outcomes described in the scientific literature are presented and the most important aspects are highlighted.

RESULTS

The vast majority of ANN applications are related to cervical cytopathology, specifically for the ANN-based, semiautomated commercial diagnostic system PAPNET. For cervical cytopathology, there is a plethora of studies relevant to the diagnostic accuracy; in addition, there are also efforts evaluating cost-effectiveness and applications on primary, secondary, or hybrid screening. For the rest of the anatomical sites, such as the gastrointestinal system, thyroid gland, urinary tract, and breast, there are significantly less efforts relevant to the application of ANNs. Additionally, there are still anatomical systems for which ANNs have never been applied on their cytological material.

CONCLUSIONS

Cytopathology is an ideal discipline to apply ANNs. In general, diagnosis is performed by experts via the light microscope. However, this approach introduces subjectivity, because this is not a universal and objective measurement process. This has resulted in the existence of a gray zone between normal and pathological cases. From the analysis of related articles, it is obvious that there is a need to perform more thorough analyses, using extensive number of cases and particularly for the nonexplored organs. Efforts to apply such systems within the laboratory test environment are required for their future uptake.

Improved Classification of Lung Cancer Using Radial Basis Function Neural Network with Affine Transforms of Voss Representation

Lung cancer is one of the diseases responsible for a large number of cancer related death cases worldwide. The recommended standard for screening and early detection of lung cancer is the low dose computed tomography. However, many patients diagnosed die within one year, which makes it essential to find alternative approaches for screening and early detection of lung cancer. We present computational methods that can be implemented in a functional multi-genomic system for classification, screening and early detection of lung cancer victims. Samples of top ten biomarker genes previously reported to have the highest frequency of lung cancer mutations and sequences of normal biomarker genes were respectively collected from the COSMIC and NCBI databases to validate the computational methods. Experiments were performed based on the combinations of Z-curve and tetrahedron affine transforms, Histogram of Oriented Gradient (HOG), Multilayer perceptron and Gaussian Radial Basis Function (RBF) neural networks to obtain an appropriate combination of computational methods to achieve improved classification of lung cancer biomarker genes. Results show that a combination of affine transforms of Voss representation, HOG genomic features and Gaussian RBF neural network perceptibly improves classification accuracy, specificity and sensitivity of lung cancer biomarker genes as well as achieving low mean square error.

Cancer Biomarkers from Genome-Scale DNA Methylation: Comparison of Evolutionary and Semantic Analysis Methods

DNA methylation profiling exploits microarray technologies, thus yielding a wealth of high-volume data. Here, an intelligent framework is applied, encompassing epidemiological genome-scale DNA methylation data produced from the Illumina’s Infinium Human Methylation 450K Bead Chip platform, in an effort to correlate interesting methylation patterns with cancer predisposition and, in particular, breast cancer and B-cell lymphoma. Feature selection and classification are employed in order to select, from an initial set of ~480,000 methylation measurements at CpG sites, predictive cancer epigenetic biomarkers and assess their classification power for discriminating healthy versus cancer related classes. Feature selection exploits evolutionary algorithms or a graph-theoretic methodology which makes use of the semantics information included in the Gene Ontology (GO) tree. The selected features, corresponding to methylation of CpG sites, attained moderate-to-high classification accuracies when imported to a series of classifiers evaluated by resampling or blindfold validation. The semantics-driven selection revealed sets of CpG sites performing similarly with evolutionary selection in the classification tasks. However, gene enrichment and pathway analysis showed that it additionally provides more descriptive sets of GO terms and KEGG pathways regarding the cancer phenotypes studied here. Results support the expediency of this methodology regarding its application in epidemiological studies.

The Application of Classification and Regression Trees for the Triage of Women for Referral to Colposcopy and the Estimation of Risk for Cervical Intraepithelial Neoplasia: A Study Based on 1625 Cases with Incomplete Data from Molecular Tests

OBJECTIVE

Nowadays numerous ancillary techniques detecting HPV DNA and mRNA compete with cytology; however no perfect test exists; in this study we evaluated classification and regression trees (CARTs) for the production of triage rules and estimate the risk for cervical intraepithelial neoplasia (CIN) in cases with ASCUS+ in cytology.

STUDY DESIGN

We used 1625 cases. In contrast to other approaches we used missing data to increase the data volume, obtain more accurate results, and simulate real conditions in the everyday practice of gynecologic clinics and laboratories. The proposed CART was based on the cytological result, HPV DNA typing, HPV mRNA detection based on NASBA and flow cytometry, p16 immunocytochemical expression, and finally age and parous status.

RESULTS

Algorithms useful for the triage of women were produced; gynecologists could apply these in conjunction with available examination results and conclude to an estimation of the risk for a woman to harbor CIN expressed as a probability.

CONCLUSIONS

The most important test was the cytological examination; however the CART handled cases with inadequate cytological outcome and increased the diagnostic accuracy by exploiting the results of ancillary techniques even if there were inadequate missing data. The CART performance was better than any other single test involved in this study.

Shirazi F. Discrimination of Human Cell Lines by Infrared Spectroscopy and Mathematical Modeling

Variations in biochemical features are extensive among cells. Identification of marker that is specific for each cell is essential for following the differentiation of stem cell and metastatic growing. Fourier transform infrared spectroscopy (FTIR) as a biochemical analysis more focused on diagnosis of cancerous cells. In this study, commercially obtained cell lines such as Human ovarian carcinoma (A2780), Human lung adenocarcinoma (A549) and Human hepatocarcinoma (HepG2) cell lines in 20 individual samples for each cell lines were used for FTIR spectral measurements. Data dimension were reduced through principal component analysis (PCA) and then subjected to neural network and linear discrimination analysis to classify FTIR pattern in different cell lines. The results showed dramatic changes of FTIR spectra among different cell types. These appeared to be associated with changes in lipid bands from CH2 symmetric and asymmetric bands, as well as amide I and amid II bands of proteins. The PCA-ANN analysis provided over 90% accuracy for classifying the spectrum of lipid section in different cell lines. This work supports future study to establish the data bank of FTIR feature for different cells and move forward to tissues as more complex systems.

Classification of lung cancer using ensemble-based feature selection and machine learning methods

Lung cancer is one of the leading causes of death worldwide. There are three major types of lung cancers, non-small cell lung cancer (NSCLC), small cell lung cancer (SCLC) and carcinoid. NSCLC is further classified into lung adenocarcinoma (LADC), squamous cell lung cancer (SQCLC) as well as large cell lung cancer. Many previous studies demonstrated that DNA methylation has emerged as potential lung cancer-specific biomarkers. However, whether there exists a set of DNA methylation markers simultaneously distinguishing such three types of lung cancers remains elusive. In the present study, ROC (Receiving Operating Curve), RFs (Random Forests) and mRMR (Maximum Relevancy and Minimum Redundancy) were proposed to capture the unbiased, informative as well as compact molecular signatures followed by machine learning methods to classify LADC, SQCLC and SCLC. As a result, a panel of 16 DNA methylation markers exhibits an ideal classification power with an accuracy of 86.54%, 84.6% and a recall 84.37%, 85.5% in the leave-one-out cross-validation (LOOCV) and independent data set test experiments, respectively. Besides, comparison results indicate that ensemble-based feature selection methods outperform individual ones when combined with the incremental feature selection (IFS) strategy in terms of the informative and compact property of features. Taken together, results obtained suggest the effectiveness of the ensemble-based feature selection approach and the possible existence of a common panel of DNA methylation markers among such three types of lung cancer tissue, which would facilitate clinical diagnosis and treatment.

Development of a multiplex methylation specific PCR suitable for (early) detection of non-small cell lung cancer

Lung cancer is a worldwide health problem and a leading cause of cancer-related deaths. Silencing of potential tumor suppressor genes (TSGs) by aberrant promoter methylation is an early event in the initiation and development of cancer. Thus, methylated cancer type-specific TSGs in DNA can serve as useful biomarkers for early cancer detection. We have now developed a "Multiplex Methylation Specific PCR" (MMSP) assay for analysis of the methylation status of multiple potential TSGs by a single PCR reaction. This method will be useful for early diagnosis and treatment outcome studies of non-small cell lung cancer (NSCLC). Genome-wide CpG methylation and expression microarrays were performed on lung cancer tissues and matched distant non-cancerous tissues from three NSCLC patients from China. Thirty-eight potential TSGs were selected and analyzed by methylation PCR on bisulfite treated DNA. On the basis of sensitivity and specificity, six marker genes, HOXA9, TBX5, PITX2, CALCA, RASSF1A, and DLEC1, were selected to establish the MMSP assay. This assay was then used to analyze lung cancer tissues and matched distant non-cancerous tissues from 70 patients with NSCLC, as well as 24 patients with benign pulmonary lesion as controls. The sensitivity of the assay was 99% (69/70). HOXA9 and TBX5 were the 2 most sensitive marker genes: 87% (61/70) and 84% (59/70), respectively. RASSF1A and DLEC1 showed the highest specificity at 99% (69/70). Using the criterion of identifying at least any two methylated marker genes, 61/70 cancer samples were positive, corresponding to a sensitivity of 87% and a specificity of 94%. Early stage I or II NSCLC could even be detected with a 100% specificity and 86% sensitivity. In conclusion, MMSP has the potential to be developed into a population-based screening tool and can be useful for early diagnosis of NSCLC. It might also be suitable for monitoring treatment outcome and recurrence.

Epigenetic reprogramming of human lung cancer cells with the extract of bovine parthenogenetic oocytes

The tumour suppressor gene silencing and proto-oncogene activation caused by epigenetic alterations plays an important role in the initiation and progression of cancer. Re-establishing the balance between the expression of tumour suppressor genes and proto-oncogenes by epigenetic modulation is a promising strategy for cancer treatment. In this study, we investigated whether cancer cells can be epigenetically reprogrammed by oocyte extract. H460 human lung cancer cells were reversibly permeabilized and incubated with the extract of bovine parthenogenetic oocytes. Bisulphite sequencing showed that bovine parthenogenetic oocyte extract induced significant demethylation at the promoters of the tumour suppressor genes RUNX3 and CDH1, but not at the promoter of the oncogenic pluripotency gene SOX2. Chromatin immunoprecipitation showed that the histone modifications at RUNX3 and CDH1 promoters were modulated towards a transcriptionally activating state, while those at SOX2 promoter towards a transcriptionally repressive state. Correspondingly, bovine parthenogenetic oocyte extract reversed the epigenetic silencing of RUNX3 and CDH1, and repressed the expression of SOX2. At the functional level, proliferation, anchorage-independent growth, migration and invasion of H460 cells was strongly inhibited. These results indicate that bovine parthenogenetic oocyte extract changes the expression patterns of tumour suppressor and oncogenic genes in cancer cells by remodelling the epigenetic modifications at their promoters. Bovine parthenogenetic oocyte extract may provide a useful tool for epigenetically reprogramming cancer cells and for dissecting the epigenetic mechanisms involved in tumorigenesis.

Prediction of lung tumor types based on protein attributes by machine learning algorithms

Early diagnosis of lung cancers and distinction between the tumor types (Small Cell Lung Cancer (SCLC) and Non-Small Cell Lung Cancer (NSCLC) are very important to increase the survival rate of patients. Herein, we propose a diagnostic system based on sequence-derived structural and physicochemical attributes of proteins that involved in both types of tumors via feature extraction, feature selection and prediction models. 1497 proteins attributes computed and important features selected by 12 attribute weighting models and finally machine learning models consist of seven SVM models, three ANN models and two NB models applied on original database and newly created ones from attribute weighting models; models accuracies calculated through 10-fold cross and wrapper validation (just for SVM algorithms). In line with our previous findings, dipeptide composition, autocorrelation and distribution descriptor were the most important protein features selected by bioinformatics tools. The algorithms performances in lung cancer tumor type prediction increased when they applied on datasets created by attribute weighting models rather than original dataset. Wrapper-Validation performed better than X-Validation; the best cancer type prediction resulted from SVM and SVM Linear models (82%). The best accuracy of ANN gained when Neural Net model applied on SVM dataset (88%). This is the first report suggesting that the combination of protein features and attribute weighting models with machine learning algorithms can be effectively used to predict the type of lung cancer tumors (SCLC and NSCLC).

Improved classification of lung cancer tumors based on structural and physicochemical properties of proteins using data mining models

Detecting divergence between oncogenic tumors plays a pivotal role in cancer diagnosis and therapy. This research work was focused on designing a computational strategy to predict the class of lung cancer tumors from the structural and physicochemical properties (1497 attributes) of protein sequences obtained from genes defined by microarray analysis. The proposed methodology involved the use of hybrid feature selection techniques (gain ratio and correlation based subset evaluators with Incremental Feature Selection) followed by Bayesian Network prediction to discriminate lung cancer tumors as Small Cell Lung Cancer (SCLC), Non-Small Cell Lung Cancer (NSCLC) and the COMMON classes. Moreover, this methodology eliminated the need for extensive data cleansing strategies on the protein properties and revealed the optimal and minimal set of features that contributed to lung cancer tumor classification with an improved accuracy compared to previous work. We also attempted to predict via supervised clustering the possible clusters in the lung tumor data. Our results revealed that supervised clustering algorithms exhibited poor performance in differentiating the lung tumor classes. Hybrid feature selection identified the distribution of solvent accessibility, polarizability and hydrophobicity as the highest ranked features with Incremental feature selection and Bayesian Network prediction generating the optimal Jack-knife cross validation accuracy of 87.6%. Precise categorization of oncogenic genes causing SCLC and NSCLC based on the structural and physicochemical properties of their protein sequences is expected to unravel the functionality of proteins that are essential in maintaining the genomic integrity of a cell and also act as an informative source for drug design, targeting essential protein properties and their composition that are found to exist in lung cancer tumors.

Identification of women for referral to colposcopy by neural networks: a preliminary study based on LBC and molecular biomarkers

Objective of this study is to investigate the potential of the learning vector quantizer neural network (LVQ-NN) classifier on various diagnostic variables used in the modern cytopathology laboratory and to build an algorithm that may facilitate the classification of individual cases. From all women included in the study, a liquid-based cytology sample was obtained; this was tested via HPV DNA test, E6/E7 HPV mRNA test, and p16 immunostaining. The data were classified by the LVQ-NN into two groups: CIN-2 or worse and CIN-1 or less. Half of the cases were used to train the LVQ-NN; the remaining cases (test set) were used for validation. Out of the 1258 cases, cytology identified correctly 72.90% of the CIN-2 or worst cases and 97.37% of the CIN-1 or less cases, with overall accuracy 94.36%. The application of the LVQ-NN on the test set allowed correct classification for 84.62% of the cases with CIN-2 or worse and 97.64% of the cases with CIN-1 or less, with overall accuracy of 96.03%. The use of the LVQ-NN with cytology and the proposed biomarkers improves significantly the correct classification of cervical precancerous lesions and/or cancer and may facilitate diagnosis and patient management.

Classification of lung cancer tumors based on structural and physicochemical properties of proteins by bioinformatics models

Rapid distinction between small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC) tumors is very important in diagnosis of this disease. Furthermore sequence-derived structural and physicochemical descriptors are very useful for machine learning prediction of protein structural and functional classes, classifying proteins and the prediction performance. Herein, in this study is the classification of lung tumors based on 1497 attributes derived from structural and physicochemical properties of protein sequences (based on genes defined by microarray analysis) investigated through a combination of attribute weighting, supervised and unsupervised clustering algorithms. Eighty percent of the weighting methods selected features such as autocorrelation, dipeptide composition and distribution of hydrophobicity as the most important protein attributes in classification of SCLC, NSCLC and COMMON classes of lung tumors. The same results were observed by most tree induction algorithms while descriptors of hydrophobicity distribution were high in protein sequences COMMON in both groups and distribution of charge in these proteins was very low; showing COMMON proteins were very hydrophobic. Furthermore, compositions of polar dipeptide in SCLC proteins were higher than NSCLC proteins. Some clustering models (alone or in combination with attribute weighting algorithms) were able to nearly classify SCLC and NSCLC proteins. Random Forest tree induction algorithm, calculated on leaves one-out and 10-fold cross validation) shows more than 86% accuracy in clustering and predicting three different lung cancer tumors. Here for the first time the application of data mining tools to effectively classify three classes of lung cancer tumors regarding the importance of dipeptide composition, autocorrelation and distribution descriptor has been reported.

Shirazi F. Patterns Prediction of Chemotherapy Sensitivity in Cancer Cell lines Using FTIR Spectrum, Neural Network and Principal Components Analysis

Drug resistance enables cancer cells to break away from cytotoxic effect of anticancer drugs. Identification of resistant phenotype is very important because it can lead to effective treatment plan. There is an interest in developing classifying models of resistance phenotype based on the multivariate data. We have investigated a vibrational spectroscopic approach in order to characterize a sensitive human ovarian cell line, A2780, and its cisplatin-resistant derivative, A2780-cp. In this study FTIR method have been evaluated via the use of principal components analysis (PCA), ANN (artificial neuronal network) and LDA (linear discriminate analysis). FTIR spectroscopy on these cells in the range of 400-4000 cm(-1) showed alteration in the secondary structure of proteins and a CH stretching vibration. We have found that the ANN models correctly classified more than 95% of the cell lines, while the LDA models with the same data sets could classify 85% of cases. In the process of different ranges of spectra, the best classification of data set in the range of 1000-2000 cm(-1) was done using ANN model, while the data set between 2500-3000 cm(-1) was more correctly classified with the LDA model. PCA of the spectral data also provide a good separation for representing the variety of cell line spectra. Our work supports the promise of ANN analysis of FTIR spectrum as a supervised powerful approach and PCA as unsupervised modeling for the development of automated methods to determine the resistant phenotype of cancer classification.

DNA hypermethylation in lung cancer is targeted at differentiation-associated genes

Aberrant DNA hypermethylation of tumor suppressor genes is thought to be an early event in tumorigenesis. Many studies have reported the methylation status of individual genes with known involvement in cancer, but an unbiased assessment of the biological function of the collective of hypermethylated genes has not been conducted so far. Based on the observation that a variety of human cancers recapitulate developmental gene expression patterns (that is activate genes normally expressed in early development and suppress late developmental genes), we hypothesized that the silencing of differentiation-associated genes in cancer could be attributed in part to DNA hypermethylation. To this end, we investigated the developmental expression patterns of genes with hypermethylated CpG islands in primary human lung carcinomas and lung cancer cell lines. We found that DNA hypermethylation primarily affects genes that are expressed in late stages of murine lung development. Gene ontology characterization of these genes shows that they are almost exclusively involved in morphogenetic differentiation processes. Our results indicate that DNA hypermethylation in cancer functions as a selective silencing mechanism of genes that are required for the maintenance of a differentiated state. The process of cellular de-differentiation that is evident on both the microscopic and transcriptional level in cancer might at least partly be mediated by these epigenetic events. Our observations provide a mechanistic explanation for induction of differentiation upon treatment with DNA methyltransferase inhibitors.

Computational Epigenetics: the new scientific paradigm

Epigenetics has recently emerged as a critical field for studying how non-gene factors can influence the traits and functions of an organism. At the core of this new wave of research is the use of computational tools that play critical roles not only in directing the selection of key experiments, but also in formulating new testable hypotheses through detailed analysis of complex genomic information that is not achievable using traditional approaches alone. Epigenomics, which combines traditional genomics with computer science, mathematics, chemistry, biochemistry and proteomics for the large-scale analysis of heritable changes in phenotype, gene function or gene expression that are not dependent on gene sequence, offers new opportunities to further our understanding of transcriptional regulation, nuclear organization, development and disease. This article examines existing computational strategies for the study of epigenetic factors. The most important databases and bioinformatic tools in this rapidly growing field have been reviewed.

Optimal search-based gene subset selection for gene array cancer classification

High dimensionality has been a major problem for gene array-based cancer classification. It is critical to identify marker genes for cancer diagnoses. We developed a framework of gene selection methods based on previous studies. This paper focuses on optimal search-based subset selection methods because they evaluate the group performance of genes and help to pinpoint global optimal set of marker genes. Notably, this paper is the first to introduce tabu search (TS) to gene selection from high-dimensional gene array data. Our comparative study of gene selection methods demonstrated the effectiveness of optimal search-based gene subset selection to identify cancer marker genes. TS was shown to be a promising tool for gene subset selection.

Epigenetic diagnostics of cancer--the application of DNA methylation markers

In recent years it has become apparent that epigenetic events are potentially equally responsible for cancer initiation and progression as genetic abnormalities. DNA methylation is the main epigenetic modification in humans. Two DNA methylation lesions coexist in human neoplasms: hypermethylation of promoter regions of specific genes within a context of genomic hypomethylation. Aberrant methylation is found at early stages of carcinogenesis and distinct types of cancer exhibit specific patterns of methylation changes. Tumor specific DNA is readily obtainable from different clinical samples and methylation status analysis often permits sensitive disease detection. Methylation markers may also serve for prognostic and predictive purposes as they often reflect the metastatic potential and sensitivity to therapy. As current findings show a great potential of recently characterised methylation markers, more studies in the field are needed in the future. Large clinical studies of newly developed markers are especially needed. The review describes the diagnostic potential of DNA methylation markers.

Survey of differentially methylated promoters in prostate cancer cell lines

DNA methylation and copy number in the genomes of three immortalized prostate epithelial and five cancer cell lines (LNCaP, PC3, PC3M, PC3M-Pro4, and PC3M-LN4) were compared using a microarray-based technique. Genomic DNA is cut with a methylation-sensitive enzyme HpaII, followed by linker ligation, polymerase chain reaction (PCR) amplification, labeling, and hybridization to an array of promoter sequences. Only those parts of the genomic DNA that have unmethylated restriction sites within a few hundred base pairs generate PCR products detectable on an array. Of 2732 promoter sequences on a test array, 504 (18.5%) showed differential hybridization between immortalized prostate epithelial and cancer cell lines. Among candidate hypermethylated genes in cancer-derived lines, there were eight (CD44, CDKN1A, ESR1, PLAU, RARB, SFN, TNFRSF6, and TSPY) previously observed in prostate cancer and 13 previously known methylation targets in other cancers (ARHI, bcl-2, BRCA1, CDKN2C, GADD45A, MTAP, PGR, SLC26A4, SPARC, SYK, TJP2, UCHL1, and WIT-1). The majority of genes that appear to be both differentially methylated and differentially regulated between prostate epithelial and cancer cell lines are novel methylation targets, including PAK6, RAD50, TLX3, PIR51, MAP2K5, INSR, FBN1, and GG2-1, representing a rich new source of candidate genes used to study the role of DNA methylation in prostate tumors.

Evidence-based medicine in pathology: an introduction

Evidence-based medicine (EBM) is a school of thought that has spread rapidly through medicine in the past 2 decades and is eliciting an increasing interest in Anatomic Pathology and Laboratory Medicine. It has been defined as "the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients." The environmental factors that created a need for EBM and basic concepts of this discipline are reviewed. Methods for the accrual and critical appraisal of the validity of available evidence and its impact, applicability and usefulness in pathology practice are discussed. Basic concepts of bayesian data analysis with an emphasis on concepts such as prior and posterior probability and the use of "holdout" or "test" data are introduced. The future of EBM in pathology is discussed and potential applications of these concepts to pathology practice and research are proposed.

The application of special technologies in diagnostic anatomic pathology: is it consistent with the principles of evidence-based medicine?

Proponents of evidence-based medicine (EBM) have emphasized the need to consider the quality of different sources of medical information and have proposed various methods to integrate available "best evidence" into rules, guidelines and other diagnostic, therapeutic and prognostic models. The various factors that can affect the internal validity of studies in anatomic pathology, such as interobserver variability, use of retrospective rather than prospective data and others, are reviewed. The need for testing for the external validity of the results of anatomic pathology studies is introduced, using "test sets" of cases that have not been used to generate the classification or prognostic models. This methodology has been seldom used in anatomic pathology to validate the generalizability of various "entities," usefulness of diagnostic tests under different conditions and other information. Basic concepts of meta-analysis for research synthesis are introduced; these methods have been seldom used in anatomic pathology to integrate information from different studies using quantitative techniques rather than summary tables that merely list the results of various publications. The potential use of decision analysis and value of information analysis for the adoption of new tests is briefly discussed.

Evidence-based medicine, medical decision analysis, and pathology

Recent advances in molecular pathology and other technologies such as proteomics present pathologists with the challenge of integrating the new information generated with high-throughput methods with current diagnostic models based mostly on histopathology and clinicopathologic correlations. Parallel developments in the field of medical informatics and bioinformatics provide the technical and mathematical methods to approach these problems in a rational manner. However, it remains unclear whether pathologists or other medical specialists will become primarily responsible for the development and maintenance of these multivariate and multidisciplinary diagnostic and prognostic models that are hoped to provide more accurate, individualized patient-based information. Evidence-based medicine (EBM) and medical decision analysis (MDA) are relatively new disciplines that use quantitative methods to assess the value of information, differentiate fact from myth, and integrate so-called best evidence into multivariate models for the assessment of prognosis, response to therapy, selection of laboratory tests, and other complex problems that influence individual patient care. We review from an epistemological viewpoint the current approach to information in pathology and describe some of the concepts developed by the practitioners of EBM and MDA.