False discovery rate

A large variation in prognosis is observed despite the use of clinical prognostic factors in patients with advanced non-small cell lung cancer (NSCLC). It is likely that this variation is due to the different biological properties of the tumour cells. In this work we aimed to identify gene signature that could predict survival in advanced NSCLC. Total RNA was extracted from five 5 μm-thick sections of the FFPE using the High Pure RNA Paraffin Kit (Roche). RNA amplification was performed using WT-Ovation™ FFPE RNA Amplification System V2 (NuGen). The amplified cDNA was then labelled and hybridised onto Illumina HumanRef-8 v3.0 Expression BeadChips. Microarray data analysis was subsequently performed using Genespring GX version 9.0. Out of 75 FFPE samples, only 32 had sufficient RNA quality and quantity for microarray gene expression analysis. Patients were grouped into long and short survival groups based on the time to cancer-related death. After normalization and filtration, 19,002 genes were selected for differential gene expression analysis. A total of 440 genes differed significantly between the long and short survival groups (ANOVA, p <; 0.05, with Benjamini and Hochberg False Discovery Rate multiple testing correction). Unsupervised Hierarchial Clustering with Pearson correlation and average linkage identified two broad clusters of patients corresponding to the long and short survival. Thirteen genes were selected based on the TTest, 2-fold expression changes, principal components analysis and univariate Cox regression analysis and risk scores were calculated for each patient. These gene signatures were independent predictors of survival. The model was validated with a published microarray data from 130 patients with NSCLC. Using Gene Set Analysis (GSA), we found certain biological processes including metastasis and chemotherapy resistance were up-regulated in the short survival group while TID pathway and MAPKKK cascade were enriched in th- - e long survival group. As the conclusion, there is several distinct gene expression profiles associated with survival of patients with advanced stage NSCLC. Survival outcomes in advanced NSCLC could be predicted based on a 13-gene signature.

DNA microarray is a technology that simultaneously evaluates quantitative measurements for the expression of thousands of genes. DNA microarrays have been used to assess gene expression between groups of cells of different organs or different populations. In order to understand the role and function of the genes, one needs the complete information about their mRNA transcripts and proteins. Unfortunately, exploring the protein functions is very difficult, due to their unique 3-dimentional complicated structure. To overcome this difficulty, one may concentrate on the mRNA molecules produced by the gene expression. In this paper, we describe some of the methods for preprocessing data for gene expression and for pairwise comparison from genomic experiments. Previous studies to assess the efficiency of different methods for pairwise comparisons have found little agreement in the lists of significant genes. Finally, we describe the procedures to control false discovery rates, sample size ...

Assessing the significance of multiple and dependent comparisons is an important, and often ignored, issue that becomes more critical as the size of data sets increases. If not accounted for, false-positive differences are very likely to be identified. The need to address this issue has led to the development of a myriad of procedures to account for multiple testing. The simplest and most widely used technique is the Bonferroni method, which controls the probability that a true null hypothesis is incorrectly rejected. However, it is a very conservative procedure. As a result, the larger the data set the greater the chances that truly significant differences will be missed. In 1995, a new criterion, the false discovery rate (FDR), was proposed to control the proportion of false declarations of significance among those individual deviations from null hypotheses considered to be significant. It is more powerful than all previously proposed methods. Multiple and dependent comparisons are also fundamental in spatial analysis. As the number of locations increases, assessing the significance of local statistics of spatial association becomes a complex matter. In this article we use empirical and simulated data to evaluate the use of the FDR approach in appraising the occurrence of clusters detected by local indicators of spatial association. Results show a significant gain in identification of meaningful clusters when controlling the FDR, in comparison to more conservative approaches. When no control is adopted, false clusters are likely to be identified. If a conservative approach is used, clusters are only partially identified and true clusters are largely missed. In contrast, when the FDR approach is adopted, clusters are fully identified. Incorporating a correction for spatial dependence to conservative methods improves the results, but not enough to match those obtained by the FDR approach.

Using a large set of high mass accuracy and resolution ETD tandem mass spectra, we characterized ETD-induced neutral losses. From these data we deduced the chemical formula for 20 of these losses. Many of them have been previously observed in electron-capture dissociation (ECD) spectra, such as losses of the side chains of arginine, aspartic acid, glutamic acid, glutamine, asparagine, leucine, histidine, and carbamidomethylated cysteine residues. With this information, we examined the diagnostic value of these amino acid-specific losses. Among 1285 peptide–spectrum matches, 92.5% have agreement between neutral loss-derived peptide amino acid composition and the peptide sequences. Moreover, we show that peptides can be uniquely identified by using only the accurate precursor mass and amino acid composition based on neutral losses; the median number of sequence candidates from an accurate mass query is reduced from 21 to 8 by adding side chain loss information. Besides increasing confidence in peptide identification, our findings suggest the potential use of these diagnostic losses in ETD spectra to improve false discovery rate estimation and to enhance the performance of scoring functions in database search algorithms.