Selecting significant genes by randomization test for cancer classification using gene expression data

Gene selection is an important task in bioinformatics studies, because the accuracy of cancer classification generally depends upon the genes that have biological relevance to the classifying problems. In this work, randomization test (RT) is used as a gene selection method for dealing with gene expression data. In the method, a statistic derived from the statistics of the regression coefficients in a series of partial least squares discriminant analysis (PLSDA) models is used to evaluate the significance of the genes. Informative genes are selected for classifying the four gene expression datasets of prostate cancer, lung cancer, leukemia and non-small cell lung cancer (NSCLC) and the rationality of the results is validated by multiple linear regression (MLR) modeling and principal component analysis (PCA). With the selected genes, satisfactory results can be obtained.     

Evaluating test statistics to select interesting genes in microarray experiments

A randomization procedure to evaluate the significance level and the false-discovery rate in complex microarray experiments is proposed. A related graph can be used to compare different test statistics that can be used to analyze the same experiment. This graph is closely related to receiver operator characteristic (ROC) curves. The proposed method is applied to a subset of the data from a cell-line experiment related to Huntington's disease. A small simulation study compares the effectiveness of the proposed procedure with the significance analysis of microarrays (SAM) procedure.     

Selecting differentially expressed genes using minimum probability of classification error

Discovery of differentially expressed genes between normal and diseased patients is a central research problem in bioinformatics. It is specially important to find few genetic markers which can be explored for diagnostic purposes. The performance of a set of markers is often measured by the associated classification accuracy. This motivates our ranking of genes depending on the minimum probability of classification errors (MPE) for each gene. In this work, we use Bayesian decision-making algorithm to compute MPE. A quantile-based probability density estimation technique is used for generating probability density functions of genes.The method is tested on three datasets: colon cancer, leukaemia, and hereditary breast cancer. The quality of the selected markers is evaluated by the classification accuracy obtained using support-vector-machine and a modified naive Bayes classifier. We obtain 96.77% accuracy in colon cancer and 97.06% accuracy in leukaemia, using only five genes in each case. Finally, using just three genes we get 100% accuracy in hereditary breast cancer.We also compare our results with those using the genes ranked by p-value and show that the genes ranked by MPE perform better or equal to those ranked by p-value.     

Influence of the test medium on azithromycin and erythromycin regression statistics

Azithromycin and erythromycin disk test results were compared to MIC values obtained in six different media. One hundred isolates were tested in triplicate, and geometric mean MICs were plotted against arithmetic mean zone diameters and regression statistics calculated. The test media evaluated did not markedly influence MIC values, but incubation in 5–7 % CO₂ resulted in a two- to four-fold decrease in the activity of both drugs. For testingHaemophilus influenzae and other species that need to be tested in 5–7 % CO₂, interpretive breakpoints for the macrolides and azalides should be modified to compensate for the anticipated decrease in activity.     

Selecting an acceptable and safe antibody detection test can present a dilemma

The Transfusion Service at Duke University Hospital has changed antibody detection methods from the use of albumin in indirect antiglobulin tests to low-ionic-strength solution (LISS), and from LISS to polyethylene glycol (PEG) in an effort to enhance the rapid detection of clinically significant antibodies. In 1996, staffing issues required the consideration of automation. Although previous studies indicated that the gel test was not as sensitive as PEG for detection of clinically significant antibodies, we chose to implement the gel test to be used with the Tecan MegaFlex-ID. We performed a retrospective analysis of identified antibodies and transfusion reactions to compare the outcomes of one year's experience with gel and PEG. We found comparable detection of potentially clinically significant antibodies by both methods and significantly fewer unwanted or clinically insignificant antibodies detected with the use of gel. Fewer delayed serologic transfusion reactions and no transfusion-associated hemolytic events occurred in the year that gel was used. Although we initially found the selection of the gel test to be a dilemma, our ultimate decision appears to have successfully protected patient safety and balanced sensitivity with specificity.     

Displacement of epsilon-proteobacterial core genes by horizontally transferred homologous genes

The introduction of novel genes by horizontal gene transfer (HGT) is considered an alternative mechanism for genetic adaptation, leading to diversification and speciation. The goal of this study was to determine which genes that are present in all sequenced epsilon-proteobacterial genomes were acquired by HGT. In our approach we used BLAST analysis to reduce the number of genes that subsequently needed to be analysed using more in-depth phylogenetic methods, including neighbour-joining and maximum likelihood. Among the 991 core genes found in all five completed epsilon-proteobacterial genome sequences, we identified 30 genes that were probably acquired by HGT. It is proposed that these genes displaced an ancestral core gene with a similar function. Although it was not possible to identify putative donor taxa for all acquired genes, it was clear that genes were acquired from a wide range of Bacteria, including Spirochaetes, Firmicutes, Actinobacteria, mycoplasmas and several subdivisions of the Proteobacteria. We did not observe HGT from Archaea to the epsilon-Proteobacteria. The majority of acquired genes were operational genes involved in transport, metabolism, signal transduction and energy production and conversion.     

Configurational statistics of polymers obtained by ring-opening polymerization of heterocyclic monomers

A general reexamination of the configurational statistics of polymers obtained by ringopening polymerization of heterocyclic monomers is presented. The polymers under study involve one or two asymmetric centres in the main chain per monomeric unit. Sequence probabilities are calculated for Bernoullian, and first- and second-order Markov propagation models. The analyses of two polymers [poly(styrene oxide) and poly(tert-butyloxirane)] for which tacticity data are available in the literature are given as examples.     

Selecting the right medical student

Medical student selection is an important but difficult task. Three recent papers by McManus et al. in BMC Medicine have re-examined the role of tests of attainment of learning (A’ levels, GCSEs, SQA) and of aptitude (AH5, UKCAT), but on a much larger scale than previously attempted. They conclude that A’ levels are still the best predictor of future success at medical school and beyond. However, A’ levels account for only 65% of the variance in performance that is found. Therefore, more work is needed to establish relevant assessment of the other 35%.Please see related research articles http://www.biomedcentral.com/1741-7015/11/242, http://www.biomedcentral.com/1741-7015/11/243 and http://www.biomedcentral.com/1741-7015/11/244.     

Selecting SNPs to identify ancestry

An individual's genotypes at a group of single-nucleotide polymorphisms (SNPs) can be used to predict that individual's ethnicity or ancestry. In medical studies, knowledge of a subject's ancestry can minimize possible confounding, and in forensic applications, such knowledge can help direct investigations. Our goal is to select a small subset of SNPs, from the millions already identified in the human genome, that can predict ancestry with a minimal error rate. The general form for this variable selection procedure is to estimate the expected error rates for sets of SNPs using a training dataset and consider those sets with the lowest error rates given their size. The quality of the estimate for the error rate determines the quality of the resulting SNPs. As the apparent error rate performs poorly when either the number of SNPs or the number of populations is large; we propose a new estimate, the Improved Bayesian Estimate. We demonstrate that selection procedures based on this estimate produce small sets of SNPs that can accurately predict ancestry. We also provide a list of the 100 optimal SNPs for identifying ancestry.