The microarray technology plays an important role of clinical oncology field. The patient can be diagnosed a symptom about cancer through microarray data. Currently, to solve classification of microarray data is still a wild open issue. Existing methods may have a good performance, but need to spend much time to analyze microarray data, such as SVM. In this thesis, we propose a novel GA-Fuzzy-based voting mechanism to find genes which affect the symptom to better diagnose patient. The proposed algorithm can blur the boundary between classes to handle the ambiguous regions. In order to simulate the gene selection mechanism, we proposed upper bound £\-Cut and lower bound £\-Cut in voting mechanism. Two groups of data collected from the literature are used to test the performance of the proposed algorithm. In the first group of dataset, experimental results show that the accuracies of five datasets using the proposed algorithm are better than those methods proposed by Pochet et al. But, there are the four datasets which the accuracies using the proposed algorithm are a little bit worse than the methods proposed by Pochet et al. For the second group of dataset, the accuracies of seven datasets using the proposed algorithm are better than KerNN proposed by Xiong and Chen. But, there are four datasets which the accuracies using the proposed algorithm are worse than KerNN proposed by Xiong and Chen. Nevertheless, experimental results show that the proposed algorithm performs the best for multi-class data.

The honey bee is one of the most familiar insects in the world, and plays an important role in the global economy providing essential pollination services to crops, fruit trees and vegetables. However, honey bee health is severely threatened by the ectoparasitic mite Varroa destructor, which feeds on the hemolymph of pupal and adult bees, resulting in loss of nutrients and circulatory fluids, decreased overall body weight and eventually the death of the bees. To investigate the molecular defense mechanisms of the honey bee against varroa mite infestation, we employed DNA microarray analysis to compare gene expression of two contrasting honey bee colony phenotypes selected from the Saskatraz breeding program. One designated as G4 is susceptible to the varroa mite, while the other designated as S88 is highly tolerant to the varroa. Total RNAs were isolated from bees at two different stages, dark-eyed pupa and adult worker, infected or non-infected with varroa mites, and used for DNA microarray analysis. The results showed that distinct sets of genes were differentially regulated in the varroa-tolerant and varroa-susceptible honey bee phenotypes, with and without varroa infestation. In both phenotypes, there were more differentially-expressed genes identified at the pupal stage than at the adult stage, indicating that at the pupal stage honey bees are more responsive to the varroa infestation than adult bees. In the phenotype comparisons, substantially more differentially-expressed genes were found in the tolerant than susceptible line, indicating that the tolerant phenotype has an increased capacity to mobilize the expression of the genes in response to varroa mite infestation. Based on function, the differentially-expressed genes could be classified into groups that are involved in olfactory signal transduction, detoxification, metabolism and exoskeleton formation, implying several possible mechanisms for the host-parasite interaction and resistance. Quantitative RT-PCR was used to confirm the data obtained from the DNA microarray hybridization. Eleven out of twelve genes selected based on the microarray data showed consistent expression patterns measured by both methods. Overall, comprehensive evaluation of the gene expression of honey bees in response to the mite infestation by DNA microarray has revealed several possible molecular mechanisms for the host defense against the pest. Identification of highly differentially expressed genes between the two phenotypes provides potential biomarkers that can be used for breeding honey bees resistant to the varroa mite.

Initial activities were focused on isolation and characterization of fruit ripening-related genes from tomato. Screening of four tomato cDNA libraries at low stringency with 10 fruit development and ripening-related genes yielded ~3000 positives clones. Microarray expression analysis of half of these positives in mature green and breaker stage fruits resulted in eight ripening-induced genes. RNA gel-blot analysis and previously published data confirmed expression for seven of the eight. One novel gene, designated LeEREBP1, was chosen for further characterization. LeEREBP1 encodes an AP2/ERF-domain transcription factor and is ethylene inducible. The expression profiles of LeEREBP1 parallel previously characterized ripening-related genes from tomato. Transgenic plants with increased and decreased expression of LeEREBP1 were generated and are currently being characterized to define the function of LeEREBP1. A large public tomato EST dataset was mined to gain insight into the tomato transcriptome. By clustering genes according to the respective expression profiles of individual tissues, tissue and developmental expression patterns were generated and genes with similar functions grouped together. Tissues effectively clustered for relatedness according to their profiles confirming the integrity of the approach used to calculate gene expression. Statistical analysis of EST prevalence in fruit and pathogenesis-related libraries resulted in 333 genes being classified as fruit ripening-induced, 185 as fruit ripening-repressed, and 169 as pathogenesis-related. We performed a parallel analysis on public EST data for grape and compared the results for ripening-induced genes to tomato to identify similar and distinct ripening factors in addition to candidates for conserved regulators of fruit ripening. An online interactive database for tomato gene expression data - Tomato Expression Database (TED) was implemented. TED contains normalized expression data for approximately 12,000 ESTs over ten time points during fruit development. It also contains comprehensive annotation of each EST. Through TED, we provide multiple approaches to pursue analysis of specific genes of interest and/or access the larger microarray dataset to identify sets of genes that may behave in a pattern of interest. In addition, a set of useful data mining and data visualization tools were developed and are under continuing expansion.

The endometrium is receptive to the embryo during a restricted window in the mid-secretory phase. My objectives were to develop a minimally-invasive endometrial sampling method for gene expression profiling, and to identify genes differentially expressed in the receptive phase. Twenty-three normo-ovulatory women underwent uterine fluid aspiration during the pre-receptive (LH+2) and receptive (LH+7) phase of the same natural cycle. RNA was extracted, reverse transcribed, amplified and hybridized to whole-genome microarrays. Unsupervised hierarchical clustering revealed self-segregation of pre-receptive and receptive samples. Importantly, profiling by uterine fluid aspiration was representative of biopsy. An unpaired t-test with a false discovery rate of 0.05 and a Δ threshold of 4-fold identified 245 unique transcripts as differentially expressed in the receptive phase. NanoString analysis validated 96% of these genes. This approach will now allow us to correlate expression of these candidate biomarkers to implantation outcomes, towards the development of clinical assays predictive for endometrial receptivity.

The endometrium is receptive to the embryo during a restricted window in the mid-secretory phase. My objectives were to develop a minimally-invasive endometrial sampling method for gene expression profiling, and to identify genes differentially expressed in the receptive phase. Twenty-three normo-ovulatory women underwent uterine fluid aspiration during the pre-receptive (LH+2) and receptive (LH+7) phase of the same natural cycle. RNA was extracted, reverse transcribed, amplified and hybridized to whole-genome microarrays. Unsupervised hierarchical clustering revealed self-segregation of pre-receptive and receptive samples. Importantly, profiling by uterine fluid aspiration was representative of biopsy. An unpaired t-test with a false discovery rate of 0.05 and a Δ threshold of 4-fold identified 245 unique transcripts as differentially expressed in the receptive phase. NanoString analysis validated 96% of these genes. This approach will now allow us to correlate expression of these candidate biomarkers to implantation outcomes, towards the development of clinical assays predictive for endometrial receptivity.

Early detection is one of the most important factors for successful treatment of cancer. Currently, scientists are searching for molecular markers that can help identify and predict outcome and chance of recurrence in patients. In this study, we demonstratet he potential impact of Y-Box binding protein-1 (YB-1) as a marker of aggressiveness and cancer recurrence in breast malignancies by screening one of the largest tissue microarrays in North America. YB-1 is an oncogenic transcription/translation factor, which is over-expressed in the majority of malignancies, including breast cancer. In the cohort of 4049 primary breast tumours, we show that YB-1 is a strong marker of aggressiveness, poor survival and cancer recurrence in all subtypes of human breast cancer with a particularly high frequency of expression in the ER negative basal-like and HER-2 breast cancer subtypes. This suggests that targeting YB-1 may provide a new avenue for therapeutic intervention in these breast cancers that are currently challenging to treat. Cox regression multivariate analysis indicates that YB-1 is second only to nodal status as a strong independent prognostic marker for poor outcome and relapse compared to established clinico-pathological biomarkers, including tumour size, age, grade, ER and HER-2 status. This finding suggests that YB-1 has great potential to be in a priority list of biomarkers for identifying the patients with a higher risk of relapse and poor outcome. Subsequently, we find an association between YB-1 and urokinase Plasminogen Activator (uPA) expression in the basal-like subtype. We then show that YB-1 is involved in the regulation of uPA expression. More importantly, silencing YB-1 or uPA results in a significant reduction in cancer cell invasion. As there are no commercially available YB-linibitors we examine the efficacy of BMS-536924, a small molecule inhibitor for activated IGF-1R/IR on SUM149 cells. We demonstrate that activated IGF-1R is associated with poor survival in primary breast tumours and, that BMS-536924 reduces uPA expression through inhibition YB-1 in SUM149 cells. We therefore conclude that YB-1 is a bio-marker for poor survival and relapse. We also indicate that YB-1 has potential use as a molecular marker in a clinical setting. Inhibiting YB-1 may provide an ideal opportunity for targeted therapy in breast cancer.

Within nutrigenomics, clustering using data generated by microarray gene expression profiles can be used to identify sub-populations of subjects that respond differently to a given diet intervention. The use of clustering analyses is promising in obesity-related research as personalized nutrition is gaining popularity. This thesis focuses on clustering a human subcutaneous adipose tissue gene expression data set obtained during a low-calorie diet intervention to aid in the prediction of 6-month weight loss maintenance. The aims of the study were (1) to identify the best performing clustering method for clustering samples, (2) to identify differential responders to the low-calorie diet, and (3) to identify the biological pathways affected during the low-calorie diet by weight maintainers and weight regainers. MCLUST performed the best when clustering samples using relative weight change and either fasting insulin or insulin resistance change. Furthermore, it identified differences in the regulation of pathways between weight maintainers and regainers.

Follicular lymphoma (FL) is an attractive model for discovering biomarkers and elucidating mechanisms of tumour progression. We hypothesized that alterations in the expression of proteins with known roles in cancer biology and hematological cells might correlate with clinical outcome and thereby shed light on biological mechanisms. Sections from a tissue microarray (TMA) containing FL samples from 67 patients were immunostained for candidate biomarkers, including p53, p16INK4a, Bcl-2, Bcl-6, MUM1, PML, phospho-ERK, and p27Kip1. The Kaplan-Meier method and log-rank test were used to identify markers that correlate significantly (p<0.05) with overall survival (OS). The chi-squared or Fisher exact test were used to examine associations between histological markers and baseline clinical features, including the Follicular Lymphoma International Prognostic Index (FLIPI) score. Expression of p16INK4a or p53, or absent CD10 expression correlated with poor survival. Patients with p16INK4a-negative tumours had a median OS of 13.4 years compared to 8.3 years for those with p16INK4a-positive tumours (p=0.006). Expression of p16INK4a was significantly associated with low hemoglobin, elevated serum lactate dehydrogenase (LDH), high histological grade, high cell proliferation index, presence of associated diffuse large B-cell lymphoma (DLBCL) and high-risk FLIPI classification. Our observation of a positive association between p16INK4a expression and indicators of tumour aggressiveness is novel and perhaps surprising since loss of the INK4a tumour suppressor gene is one of the most frequently observed lesions in human cancers, including lymphoma. Expression of p16INK4a may be part of a cellular response to unidentified pro-mitotic mutations, such as deleterious mutations of the RB tumour suppressor gene, associated with more aggressive instances of FL. Immunostaining FL diagnostic biopsies for expression of p16INK4a may serve as an informative prognostic biomarker to aid clinicians managing FL patients. / Thesis (Master, Pathology & Molecular Medicine) -- Queen's University, 2008-07-04 15:55:16.121

RNA molecules isolated from FFPE samples are highly fragmented and modified, and generally deemed unsuitable for downstream gene expression profiling. With the development of molecular biology, there has been growing interest in profiling archival FFPE samples. Successful profiling of transcripts from FFPE samples would greatly expand tissue sources for large scale gene expression studies; also it would pave the way for future applications on the type of tissue readily available in the clinical setting. So far, there is a lack of systemic studies evaluating the quality of RNA isolated from routinely processed FFPE samples, and it has remained difficult to assess how well FFPE-derived RNA mirrors the status of RNA isolated before fixation. In this project, the similarity of miRNA and mRNA profiles between matched frozen and FFPE lymphoid hyperplasia tissues (N=7 for miRNA comparison, N=4 for mRNA comparison) were evaluated. We found consistently good correlation (mean of Pearson coefficient=0.939, mean of Spearman coefficient=0.905, mean of Kendall tau=0.744) between matched frozen and FFPE-derived miRNA profiles, suggesting FFPE samples may retain miRNA expression information quite well. This has major positive implications for research using FFPE samples, as miRNA profiling becomes more prominent in bioprofiling studies. On the contrary, mRNA isolated from FFPE samples showed less correlation (Spearman coefficient less than 0.75) with its frozen counterpart on the Agilent microarray platform. With a post extraction heat treatment aimed at reversing base modifications and cross linking structures, obvious global mRNA quality improvement was observed in cases where samples appeared to be heavily cross linked, but was less effective and even detrimental in cases where cross linking was less prominent. This research suggests that the extent of cross linking may be critical in terms of determining whether a particular FFPE tissue will become a useful source of mRNA for global profiling studies / Thesis (Master, Pathology & Molecular Medicine) -- Queen's University, 2008-09-26 10:49:50.044

No description available.