Biomedical subjects

Inge Jonassen

Publications and source records attributed to Inge Jonassen.

At least 19 recordsLinked to original sources

PubMed2007

ERG upregulation and related ETS transcription factors in prostate cancer.

The aim of this study was to identify and validate differentially expressed genes in matched pairs of benign and malignant prostate tissue. Samples included 29 histologically verified primary tumors and 23 benign controls. Microarray analysis was initially performed using a sequence verified set of 40,000 human cDNA clones. Among the genes most consistently and highly upregulated in prostate cancer was the ETS family transcription factor ERG (ETS related gene). This finding was validated in an expanded patient series (37 tumors and 38 benign samples) using DNA oligonucleotide microarray and real-time quantitative PCR assays. ERG was 20- to more than 100-fold overexpressed in prostate cancer compared with benign prostate tissue in more than 50% of patients according to quantitative PCR. Surprisingly, ERG mRNA levels were found to be significantly higher in the endothelial cell line, HUVEC, than in the prostate cell lines PC3, DU145 and LNCaP. In situ hybridization of prostate cancer tissue revealed that ERG was abundantly expressed in both prostate cancer cells and associated endothelial cells. The consistency and magnitude of ERG overexpression in prostate cancer appeared unique, but several related ETS transcription factors were also overexpressed in matched pairs of tumor and benign samples, whereas ETS2 was significantly underexpressed. Our findings support the hypothesis that ERG overexpression and related ETS transcription factors are important for early prostate carcinogenesis.

DNA Primers↗

PubMed2006

Angiogenesis-independent tumor growth mediated by stem-like cancer cells.

In this work, highly infiltrative brain tumors with a stem-like phenotype were established by xenotransplantation of human brain tumors in immunodeficient nude rats. These tumors coopted the host vasculature and presented as an aggressive disease without signs of angiogenesis. The malignant cells expressed neural stem cell markers, showed a migratory behavior similar to normal human neural stem cells, and gave rise to tumors in vivo after regrafting. Serial passages in animals gradually transformed the tumors into an angiogenesis-dependent phenotype. This process was characterized by a reduction in stem cells markers. Gene expression profiling combined with high throughput immunoblotting analyses of the angiogenic and nonangiogenic tumors identified distinct signaling networks in the two phenotypes. Furthermore, proinvasive genes were up-regulated and angiogenesis signaling genes were down-regulated in the stem-like tumors. In contrast, proinvasive genes were down-regulated in the angiogenesis-dependent tumors derived from the stem-like tumors. The described angiogenesis-independent tumor growth and the uncoupling of invasion and angiogenesis, represented by the stem-like cancer cells and the cells derived from them, respectively, point at two completely independent mechanisms that drive tumor progression. This article underlines the need for developing therapies that specifically target the stem-like cell pools in tumors.

Animals↗

PubMed2006

Protein model refinement using structural fragment tessellation.

A method is described for the refinement of rough protein models based on finding a selection of structural fragments that match the model. Unlike most fragment-based methods, these are not necessarily contiguous in the sequence and form a tiling (tessellation) that covers most of the structure. The residue positions of the fragments are then used as a target for the model atoms to generate a revised model which is used as the basis of a subsequent pattern definition and search. The method was shown to improve the recognition of the native fold in a series of decoys largely as a result of improved secondary structure representation.

Algorithms↗

PubMed2006

Dynamic domain threading.

A modeling method is described that avoids the need to consider the domain structure of the template used for modeling, and automatically extracts compact fragments of structure that would be of a suitable size to build the model. This aids automation as the size or nature of the template structure can be ignored and does not have to be broken into domain (or multi-domain) units beforehand. The approach leads to the generation of a large number of models each based on slightly differing domain definitions and this variation was further increased by considering alternative secondary structure predictions. Each model, of which there may be thousands, takes the form of a complete alpha-carbon trace and some methods (including residue burial) were investigated for their power to discriminate good models from bad models using decoys. The method is also compared to an earlier retroviral capsid modeling problem for which the X-ray structure is now known. Some potential extensions of the approach to more distant modeling problems are discussed.

Algorithms↗

PubMed2006

RBR: library-less repeat detection for ESTs.

MOTIVATION: Repeat sequences in ESTs are a source of problems, in particular for clustering. ESTs are therefore commonly masked against a library of known repeats. High quality repeat libraries are available for the widely studied organisms, but for most other organisms the lack of such libraries is likely to compromise the quality of EST analysis. RESULTS: We present a fast, flexible and library-less method for masking repeats in EST sequences, based on match statistics within the EST collection. The method is not linked to a particular clustering algorithm. Extensive testing on datasets using different clustering methods and a genomic mapping as reference shows that this method gives results that are better than or as good as those obtained using RepeatMasker with a repeat library. AVAILABILITY: The implementation of RBR is available under the terms of the GPL from http://www.ii.uib.no/~ketil/bioinformatics CONTACT: ketil.malde@bccs.uib.no SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Algorithms↗

PubMed2006

Gene expression profiles of head and neck carcinomas from Sudanese and Norwegian patients reveal common biological pathways regardless of race and lifestyle.

PURPOSE: To explore possible range of gene expression profiles in head and neck squamous cell carcinomas (HNSCC) and pairwised normal controls from Sudanese (n = 72) and Norwegian (n = 45) patients using a 15K cDNA microarray and to correlate the findings with clinicopathologic variables. EXPERIMENTAL DESIGN: Samples from Sudan were grouped according to anatomic location/patients' habit of toombak (snuff) use, and 37 pools of 2 to 11 tumors matched to 37 pools of their normal controls from the same patients, respectively, were prepared. For Norway, eight pools of 3 to 11 tumors matched to eight pools of their normal controls from the same patients, respectively, were prepared according to anatomic location. Pools (n = 45) were hybridized to microarrays. For controls, 33 of the pools were hybridized against Human Reference RNA. Scanned array images were recorded, and data analysis was done in groups. For verification, results for selected genes were analyzed using quantitative real-time PCR/immunohistochemistry. RESULTS: We identified 136 genes from Sudan and 154 from Norway as differentially expressed between tumors and controls. Changes of the genes found were confirmed in >70% of the pools by hybridization against Reference RNA. Seventy-three genes and three main pathways (signal transduction, cell communication, and ligand-receptor interaction) were of relevance to the HNSCCs from both countries. Hierarchical clustering of the 73 genes identified subclasses of mixed tumors from the two populations, two independent subgroups for Norwegian tumors by their anatomic sites, and five subgroups for Sudanese tumors by their toombak habits. Quantitative real-time PCR/immunohistochemistry validated the microarray-based data. CONCLUSIONS: Differences in gene expression between tumor and nontumor tissues were identified in HNSCCs. Analysis of the two population groups revealed a common set of 73 genes within three main biological pathways. This indicates that the development of HNSCCs is mediated by similar biological pathways regardless of differences related to race, ethnicity, lifestyle, and/or exposure to environmental carcinogens. Of particular interest, however, was the valuable association of gene expression signature found with toombak use and anatomic site of the tumors.

Aged↗

PubMed2006

cDNA microarray analysis of non-selected cases of acute myeloid leukemia demonstrates distinct clustering independent of cytogenetic aberrations and consistent with morphological signs of differentiation.

Acute myeloid leukemia (AML) is a heterogeneous disease with respect to biology and clinical course. Until now the basis for prognostic evaluation and therapeutic decision has been the karyotype, genetic FLT3 abnormalities and the initial chemotherapy response. A question that has emerged is if extensive gene expression analysis may supplement or partly replace current diagnostics. In an attempt to address this question, we performed cDNA microarray analysis on peripheral blood samples of 25 patients with newly diagnosed AML with high blast counts. The patients were randomly selected from a large group of consecutive patients. Leave-one-out crossvalidation (LOOCV) showed with high accuracy that gene expression classifiers could predict if leukaemia samples belonged to the FAB AML-M1 or to the FAB AML-M2 groups. An unsupervised two-dimensional hierarchical cluster analysis generated 3 patient subgroups. Except for an accumulation of samples classified as FAB M1 and M2 in cluster 3, there was no evident relationship between the clusters and the FAB classification. Each subgroup displayed clearly distinguished gene expression patterns validated using real-time quantitative PCR analysis. The identification of specific gene expressions that together constitute regulatory modules must complement cluster analyses in order to achieve an accurate basis for prognosis and prediction.

Adult↗

PubMed2005

Masking repeats while clustering ESTs.

A problem in EST clustering is the presence of repeat sequences. To avoid false matches, repeats have to be masked. This can be a time-consuming process, and it depends on available repeat libraries. We present a fast and effective method that aims to eliminate the problems repeats cause in the process of clustering. Unlike traditional methods, repeats are inferred directly from the EST data, we do not rely on any external library of known repeats. This makes the method especially suitable for analysing the ESTs from organisms without good repeat libraries. We demonstrate that the result is very similar to performing standard repeat masking before clustering.

Cluster Analysis↗

PubMed2005

Gene expression profiling of minor salivary glands clearly distinguishes primary Sjögren's syndrome patients from healthy control subjects.

OBJECTIVE: To identify gene expression signatures in minor salivary glands (MSGs) from patients with primary Sjogren's syndrome (SS). METHODS: A 16K complementary DNA microarray was used to generate gene expression profiles in MSGs obtained from 10 patients with primary SS and 10 control subjects. The data were analyzed by 2 different strategies, one strict primary analysis and one subanalysis that allowed for inclusion of genes with no signal in more than 3 samples from each group. The results were validated by quantitative reverse transcriptase-polymerase chain reaction techniques. RESULTS: We found a distinct difference in gene expression levels in MSGs, enabling a simple class prediction method to correctly classify 19 of the 20 samples as either patient or control, based on the top 5 differentially expressed genes. The 50 most differentially expressed genes in the primary SS group compared with the control group were all up-regulated, and a clear pattern of genes involved in chronic inflammation was found. CXCL13 and CD3D were expressed in >/=90% of primary SS patients and in </=10% of the controls. Lymphotoxin beta, as well as a number of major histocompatibility complex genes, cytokines, and lymphocyte activation factors, manifested its role in the pathogenesis of SS. Numerous type I interferon genes related to virus infection were found among the top 200 genes, with increased expression in primary SS. Interestingly, the expression of carbonic anhydrase II, which is essential in saliva production and secretion, and the apoptosis regulator Bcl-2-like 2 were down-regulated in primary SS patients. CONCLUSION: We have identified distinct gene expression profiles in MSGs from patients with primary SS that provide new knowledge about groups of genes that are up-regulated or down-regulated during disease, constituting an excellent platform for forthcoming functional studies.

Adult↗

PubMed2005

CD34 expression in native human acute myelogenous leukemia blasts: differences in CD34 membrane molecule expression are associated with different gene expression profiles.

BACKGROUND: The stem cell marker CD34 is expressed by leukemia blasts only for a subset of patients with acute myelogenous leukemia (AML). It is still controversial as to whether CD34 expression (defined as at least 10-20% positive cells) has any prognostic effect in patients with AML who receive intensive chemotherapy. The present study investigated whether gene expression profiling could be used to further subclassify CD34(+) AML cell populations. METHODS: AML blasts derived from 25 patients were examined; these patients were randomly selected from a larger consecutive group of patients. CD34 protein expression was determined by flow cytometry and expressed as the percentage of positive cells. Gene expression profiles were determined by complementary DNA microarrays. RESULTS: By unsupervised hierarchical clustering our patients could be grouped into two or three major subsets depending on the methodologic approach before clustering analysis (filtering or flooring of data, respectively). However, both approaches identified a cluster characterized by high gene expression and membrane molecule level of CD34. When using the floored expression profiles, the patient cluster characterized by increased CD34 gene expression was also characterized by a high percentage of CD34(+) cells (median 82%, range 56-100%) compared with the two other major clusters (median 19%, range <1-55%), but three of four outpatients also showed a high percentage of CD34(+) cells. CONCLUSION: A major proportion of patients with AML and high CD34 expression (usually >80% CD34(+) cells; nearly all patients had >50% positive cells) showed similarities in gene expression profile. In contrast, patients with lower CD34 expression often had a profile similar to those of patients regarded as CD34(-) according to conventional criteria. Our results suggest that the possible prognostic effect of CD34 expression should be reevaluated in clinical studies using additional or alternative cutoff values to describe CD34 expression.

Adult↗

PubMed2005

Gene expression profiles in prostate cancer: association with patient subgroups and tumour differentiation.

Prostate carcinoma is the most common cancer of western men and is a markedly heterogeneous disease. The aim of this study was to identify signatures of differentially expressed genes in prostate cancer using DNA microarray technology, evaluating expression profiles in matched pairs of benign and malignant tissue. Samples were collected from 33 radical prostatectomies, and 52 specimens were included, representing 29 histologically verified primary tumours, 19 paired samples of malignant and benign tissue, and 4 non-paired benign tissue samples. Microarray analysis was performed using an expanded sequence verified set of 40,000 human cDNA clones, revealing several genes with significant differences between malignant and benign tissue, including recently reported genes like alpha-methylacyl-CoA racemase (AMACR) and hepsin, as well as genes relevant for tumour development and progression. Leave out cross validation (LOCV) test correctly predicted tumour or benign tissue in 47 (90.3%) out of 52 cases, significantly better than cross validation tests using randomly permuted tissue labels. Unsupervised clustering analysis revealed 3 distinct patient clusters significantly associated with Gleason score, and high grade tumours (Gleason score >/=7) accumulated in cluster 1 (C1). Gene expression profiles correctly predicted 100% of tumour samples segregating to C1, as also validated by LOCV. Gene expression profiles were analysed in filtered and floored datasets with similar results, and a pair-wise design was also tested. Gene expression profiles provided tumour clusters linked to differentiation, and revealed novel markers relevant for molecular classification, grading and therapy of prostate cancer.

Cluster Analysis↗

PubMed2004

A graph based algorithm for generating EST consensus sequences.

MOTIVATION: EST sequences constitute an abundant, yet error prone resource for computational biology. Expressed sequences are important in gene discovery and identification, and they are also crucial for the discovery and classification of alternative splicing. An important challenge when processing EST sequences is the reconstruction of mRNA by assembling EST clusters into consensus sequences. RESULTS: In contrast to the more established assembly tools, we propose an algorithm that constructs a graph over sequence fragments of fixed size, and produces consensus sequences as traversals of this graph. We provide a tool implementing this algorithm, and perform an experiment where the consensus sequences produced by our implementation, as well as by currently available tools, are compared to mRNA. The results show that our proposed algorithm in a majority of the cases produces consensus of higher quality than the established sequence assemblers and at a competitive speed. AVAILABILITY: The source code for the implementation is available under a GPL license from http://www.ii.uib.no/~ketil/bioinformatics/ CONTACT: ketil@ii.uib.no.

Algorithms↗

PubMed2004

XHM: a system for detection of potential cross hybridizations in DNA microarrays.

BACKGROUND: Microarrays have emerged as the preferred platform for high throughput gene expression analysis. Cross-hybridization among genes with high sequence similarities can be a source of error reducing the reliability of DNA microarray results. RESULTS: We have developed a tool called XHM (cross hybridization on microarrays) for assessment of the reliability of hybridization signals by detecting potential cross-hybridizations on DNA microarrays. This is done by comparing the sequences of the probes against an extensive database representing the transcriptome of the organism in question. XHM is available online at http://www.bioinfo.no/tools/xhm/. CONCLUSIONS: Using XHM with its user-adjustable parameters will enable scientists to check their lists of differentially expressed genes from microarray experiments for potential cross-hybridizations. This provides information that may be useful in the validation of the microarray results.

Animals↗

PubMed2004

A structural pattern-based method for protein fold recognition.

A method (SPREK) was developed to evaluate the register of a sequence on a structure based on the matching of structural patterns against a library derived from the protein structure databank. The scores obtained were normalized against random background distributions derived from sequence shuffling and permutation methods. 'Random' structures were also used to evaluate the effectiveness of the method. These were generated by a simple random-walk and a more sophisticated structure prediction method that produced protein-like folds. For comparison with other methods, the performance of the method was assessed using collections of models including decoys and models from the CASP-5 exercise. The performance of SPREK on the decoy models was equivalent to (and sometimes better than) those obtained with more complex approaches. An exception was the two smallest proteins, for which SPREK did not perform well due to a lack of patterns. Using the best parameter combination from trials on decoy models, the CASP models of intermediate difficulty were evaluated by SPREK and the quality of the top scoring model was evaluated by its CASP ranking. Of the 14 targets in this class, half lie in the top 10% (out of around 140 models for each target). The two worst rankings resulted from the selection by our method of a well-packed model that was based on the wrong fold. Of the other poor rankings, one was the smallest protein and the others were the four largest (all over 250 residues).

Algorithms↗

PubMed2004

LSimpute: accurate estimation of missing values in microarray data with least squares methods.

Microarray experiments generate data sets with information on the expression levels of thousands of genes in a set of biological samples. Unfortunately, such experiments often produce multiple missing expression values, normally due to various experimental problems. As many algorithms for gene expression analysis require a complete data matrix as input, the missing values have to be estimated in order to analyze the available data. Alternatively, genes and arrays can be removed until no missing values remain. However, for genes or arrays with only a small number of missing values, it is desirable to impute those values. For the subsequent analysis to be as informative as possible, it is essential that the estimates for the missing gene expression values are accurate. A small amount of badly estimated missing values in the data might be enough for clustering methods, such as hierachical clustering or K-means clustering, to produce misleading results. Thus, accurate methods for missing value estimation are needed. We present novel methods for estimation of missing values in microarray data sets that are based on the least squares principle, and that utilize correlations between both genes and arrays. For this set of methods, we use the common reference name LSimpute. We compare the estimation accuracy of our methods with the widely used KNNimpute on three complete data matrices from public data sets by randomly knocking out data (labeling as missing). From these tests, we conclude that our LSimpute methods produce estimates that consistently are more accurate than those obtained using KNNimpute. Additionally, we examine a more classic approach to missing value estimation based on expectation maximization (EM). We refer to our EM implementations as EMimpute, and the estimate errors using the EMimpute methods are compared with those our novel methods produce. The results indicate that on average, the estimates from our best performing LSimpute method are at least as accurate as those from the best EMimpute algorithm.

Algorithms↗

PubMed2003

Tumor classification and marker gene prediction by feature selection and fuzzy c-means clustering using microarray data.

BACKGROUND: Using DNA microarrays, we have developed two novel models for tumor classification and target gene prediction. First, gene expression profiles are summarized by optimally selected Self-Organizing Maps (SOMs), followed by tumor sample classification by Fuzzy C-means clustering. Then, the prediction of marker genes is accomplished by either manual feature selection (visualizing the weighted/mean SOM component plane) or automatic feature selection (by pair-wise Fisher's linear discriminant). RESULTS: The proposed models were tested on four published datasets: (1) Leukemia (2) Colon cancer (3) Brain tumors and (4) NCI cancer cell lines. The models gave class prediction with markedly reduced error rates compared to other class prediction approaches, and the importance of feature selection on microarray data analysis was also emphasized. CONCLUSIONS: Our models identify marker genes with predictive potential, often better than other available methods in the literature. The models are potentially useful for medical diagnostics and may reveal some insights into cancer classification. Additionally, we illustrated two limitations in tumor classification from microarray data related to the biology underlying the data, in terms of (1) the class size of data, and (2) the internal structure of classes. These limitations are not specific for the classification models used.

Biomarkers, Tumor↗

PubMed2003

Fast sequence clustering using a suffix array algorithm.

MOTIVATION: Efficient clustering is important for handling the large amount of available EST sequences. Most contemporary methods are based on some kind of all-against-all comparison, resulting in a quadratic time complexity. A different approach is needed to keep up with the rapid growth of EST data. RESULTS: A new, fast EST clustering algorithm is presented. Sub-quadratic time complexity is achieved by using an algorithm based on suffix arrays. A prototype implementation has been developed and run on a benchmark data set. The produced clusterings are validated by comparing them to clusterings produced by other methods, and the results are quite promising. AVAILABILITY: The source code for the prototype implementation is available under a GPL license from http://www.ii.uib.no/~ketil/bio/.

Algorithms↗

PubMed2002

New feature subset selection procedures for classification of expression profiles.

BACKGROUND: Methods for extracting useful information from the datasets produced by microarray experiments are at present of much interest. Here we present new methods for finding gene sets that are well suited for distinguishing experiment classes, such as healthy versus diseased tissues. Our methods are based on evaluating genes in pairs and evaluating how well a pair in combination distinguishes two experiment classes. We tested the ability of our pair-based methods to select gene sets that generalize the differences between experiment classes and compared the performance relative to two standard methods. To assess the ability to generalize class differences, we studied how well the gene sets we select are suited for learning a classifier. RESULTS: We show that the gene sets selected by our methods outperform the standard methods, in some cases by a large margin, in terms of cross-validation prediction accuracy of the learned classifier. We show that on two public datasets, accurate diagnoses can be made using only 15-30 genes. Our results have implications for how to select marker genes and how many gene measurements are needed for diagnostic purposes. CONCLUSION: When looking for differential expression between experiment classes, it may not be sufficient to look at each gene in a separate universe. Evaluating combinations of genes reveals interesting information that will not be discovered otherwise. Our results show that class prediction can be improved by taking advantage of this extra information.

Colon↗