Comparison of five popular normalization methods using a spike-in proteomics data set

Mass spectrometry (MS)-based proteomics has seen significant technical advances during the past two decades and mass spectrometry has become a central tool in many biosciences. Despite the popularity of MS-based methods, the handling of the systematic non-biological variation in the data remains a common problem. This biasing variation can result from several sources ranging from sample handling to differences caused by the instrumentation. Normalization is the procedure which aims to account for this biasing variation and make samples comparable. Many normalization methods commonly used in proteomics have been adapted from the DNA-microarray world. Studies comparing normalization methods with proteomics data sets using some variability measures exist. However, a more thorough comparison looking at the quantitative and qualitative differences of the performance of the different normalization methods and at their ability in preserving the true differential expression signal of proteins, is lacking. In this thesis, several popular and widely used normalization methods (the Linear regression normalization, Local regression normalization, Variance stabilizing normalization, Quantile-normalization, Median central tendency normalization and also variants of some of the forementioned methods), representing different strategies in normalization are being compared and evaluated with a benchmark spike-in proteomics data set. The normalization methods are evaluated in several ways. The performance of the normalization methods is evaluated qualitatively and quantitatively on a global scale and in pairwise comparisons of sample groups. In addition, it is investigated, whether performing the normalization globally on the whole data or pairwise for the comparison pairs examined, affects the performance of the normalization method in normalizing the data and preserving the true differential expression signal. In this thesis, both major and minor differences in the performance of the different normalization methods were found. Also, the way in which the normalization was performed (global normalization of the whole data or pairwise normalization of the comparison pair) affected the performance of some of the methods in pairwise comparisons. Differences among variants of the same methods were also observed.

ArrayMining: a modular web-application for microarray analysis combining ensemble and consensus methods with cross-study normalization

Background: Statistical analysis of DNA microarray data provides a valuable diagnostic tool for the investigation of genetic components of diseases. To take advantage of the multitude of available data sets and analysis methods, it is desirable to combine both different algorithms and data from different studies. Applying ensemble learning, consensus clustering and cross-study normalization methods for this purpose in an almost fully automated process and linking different analysis modules together under a single interface would simplify many microarray analysis tasks. Results: We present ArrayMining.net, a web-application for microarray analysis that provides easy access to a wide choice of feature selection, clustering, prediction, gene set analysis and cross-study normalization methods. In contrast to other microarray-related web-tools, multiple algorithms and data sets for an analysis task can be combined using ensemble feature selection, ensemble prediction, consensus clustering and cross-platform data integration. By interlinking different analysis tools in a modular fashion, new exploratory routes become available, e.g. ensemble sample classification using features obtained from a gene set analysis and data from multiple studies. The analysis is further simplified by automatic parameter selection mechanisms and linkage to web tools and databases for functional annotation and literature mining. Conclusion: ArrayMining.net is a free web-application for microarray analysis combining a broad choice of algorithms based on ensemble and consensus methods, using automatic parameter selection and integration with annotation databases.

Arrow plot and CA maps on microarray preprocessing methods

Microarray allow to monitoring simultaneously thousands of genes, where the abundance of the transcripts under a same experimental condition at the same time can be quantified. Among various available array technologies, double channel cDNA microarray experiments have arisen in numerous technical protocols associated to genomic studies, which is the focus of this work. Microarray experiments involve many steps and each one can affect the quality of raw data. Background correction and normalization are preprocessing techniques to clean and correct the raw data when undesirable fluctuations arise from technical factors. Several recent studies showed that there is no preprocessing strategy that outperforms others in all circumstances and thus it seems difficult to provide general recommendations. In this work, it is proposed to use exploratory techniques to visualize the effects of preprocessing methods on statistical analysis of cancer two-channel microarray data sets, where the cancer types (classes) are known. For selecting differential expressed genes the arrow plot was used and the graph of profiles resultant from the correspondence analysis for visualizing the results. It was used 6 background methods and 6 normalization methods, performing 36 pre-processing methods and it was analyzed in a published cDNA microarray database (Liver) available at http://genome-www5.stanford.edu/ which microarrays were already classified by cancer type. All statistical analyses were performed using the R statistical software.

Impact of normalization on miRNA microarray expression profiling.

Profiling miRNA levels in cells with miRNA microarrays is becoming a widely used technique. Although normalization methods for mRNA gene expression arrays are well established, miRNA array normalization has so far not been investigated in detail. In this study we investigate the impact of normalization on data generated with the Agilent miRNA array platform. We have developed a method to select nonchanging miRNAs (invariants) and use them to compute linear regression normalization coefficients or variance stabilizing normalization (VSN) parameters. We compared the invariants normalization to normalization by scaling, quantile, and VSN with default parameters as well as to no normalization using samples with strong differential expression of miRNAs (heart-brain comparison) and samples where only a few miRNAs are affected (by p53 overexpression in squamous carcinoma cells versus control). All normalization methods performed better than no normalization. Normalization procedures based on the set of invariants and quantile were the most robust over all experimental conditions tested. Our method of invariant selection and normalization is not limited to Agilent miRNA arrays and can be applied to other data sets including those from one color miRNA microarray platforms, focused gene expression arrays, and gene expression analysis using quantitative PCR.

Quantifying ChIP-seq data: a spiking method providing an internal reference for sample-to-sample normalization.

Chromatin immunoprecipitation followed by deep sequencing (ChIP-seq) experiments are widely used to determine, within entire genomes, the occupancy sites of any protein of interest, including, for example, transcription factors, RNA polymerases, or histones with or without various modifications. In addition to allowing the determination of occupancy sites within one cell type and under one condition, this method allows, in principle, the establishment and comparison of occupancy maps in various cell types, tissues, and conditions. Such comparisons require, however, that samples be normalized. Widely used normalization methods that include a quantile normalization step perform well when factor occupancy varies at a subset of sites, but may miss uniform genome-wide increases or decreases in site occupancy. We describe a spike adjustment procedure (SAP) that, unlike commonly used normalization methods intervening at the analysis stage, entails an experimental step prior to immunoprecipitation. A constant, low amount from a single batch of chromatin of a foreign genome is added to the experimental chromatin. This "spike" chromatin then serves as an internal control to which the experimental signals can be adjusted. We show that the method improves similarity between replicates and reveals biological differences including global and largely uniform changes.

Domain Adaptation in remote sensing: increasing the portability of land-cover classifiers

Among the types of remote sensing acquisitions, optical images are certainly one of the most widely relied upon data sources for Earth observation. They provide detailed measurements of the electromagnetic radiation reflected or emitted by each pixel in the scene. Through a process termed supervised land-cover classification, this allows to automatically yet accurately distinguish objects at the surface of our planet. In this respect, when producing a land-cover map of the surveyed area, the availability of training examples representative of each thematic class is crucial for the success of the classification procedure. However, in real applications, due to several constraints on the sample collection process, labeled pixels are usually scarce. When analyzing an image for which those key samples are unavailable, a viable solution consists in resorting to the ground truth data of other previously acquired images. This option is attractive but several factors such as atmospheric, ground and acquisition conditions can cause radiometric differences between the images, hindering therefore the transfer of knowledge from one image to another. The goal of this Thesis is to supply remote sensing image analysts with suitable processing techniques to ensure a robust portability of the classification models across different images. The ultimate purpose is to map the land-cover classes over large spatial and temporal extents with minimal ground information. To overcome, or simply quantify, the observed shifts in the statistical distribution of the spectra of the materials, we study four approaches issued from the field of machine learning. First, we propose a strategy to intelligently sample the image of interest to collect the labels only in correspondence of the most useful pixels. This iterative routine is based on a constant evaluation of the pertinence to the new image of the initial training data actually belonging to a different image. Second, an approach to reduce the radiometric differences among the images by projecting the respective pixels in a common new data space is presented. We analyze a kernel-based feature extraction framework suited for such problems, showing that, after this relative normalization, the cross-image generalization abilities of a classifier are highly increased. Third, we test a new data-driven measure of distance between probability distributions to assess the distortions caused by differences in the acquisition geometry affecting series of multi-angle images. Also, we gauge the portability of classification models through the sequences. In both exercises, the efficacy of classic physically- and statistically-based normalization methods is discussed. Finally, we explore a new family of approaches based on sparse representations of the samples to reciprocally convert the data space of two images. The projection function bridging the images allows a synthesis of new pixels with more similar characteristics ultimately facilitating the land-cover mapping across images.

Expression study of the atherosclerosis mouse aorta reveals significant disturbance of calcium signaling pathway

Atherosclerosis is a complex disease resulting from interactions of genetic and environmental risk factors leading to heart failure and stroke. Using an atherosclerotic mouse model (ldlr-/-, apobec1-/- designated as LDb), we performed microarray analysis to identify candidate genes and pathways, which are most perturbed in changes in the following risk factors: genetics (control C57BL/6 vs. LDb mice), shearstress (lesion-prone vs. lesion-resistant regions in LDb mice), diet (chow vs. high fat fed LDb mice) and age (2-month-old vs. 8-month old LDb mice). ^ Atherosclerotic lesion quantification and lipid profile studies were performed to assess the disease phenotype. A microarray study was performed on lesion-prone and lesion-resistant regions of each aorta. Briefly, 32 male C57BL/6 and LDb mice (n =16/each) were fed on either chow or high fat diet, sacrificed at 2- and 8-months old, and RNA isolated from the aortic lesion-prone and aortic lesion-resistant segments. Using 64 Affymetrix Murine 430 2.0 chips, we profiled differentially expressed genes with the cut off value of FDR ≤ 0.15 for t-test, and q <0.0001 for the ANOVA. The data were normalized using two normalization methods---invariant probe sets (Loess) and Quantile normalization, the statistical analysis was performed using t-tests and ANOVA, and pathway characterization was done using Pathway Express (Wayne State). The result identified the calcium signaling pathway as the most significant overrepresented pathway, followed by focal adhesion. In the calcium signaling pathway, 56 genes were found to be significantly differentially expressed out of 180 genes listed in the KEGG calcium signaling pathway. Nineteen of these genes were consistently identified by both statistical tests, 11 of which were unique to the test, and 26 were unique to the ANOVA test, using the cutoffs noted above. ^ In conclusion, this finding suggested that hypercholesterolemia drives the disease progression by altering the expression of calcium channels and regulators which subsequently results in cell differentiation, growth, adhesion, cytoskeletal change and death. Clinically, this pathway may serve as an important target for future therapeutic intervention, and thus the calcium signaling pathway may serve as an important target for future diagnostic and therapeutic intervention. ^

Time series distances measures to analyze in-air signatures to authenticate users on mobile phones

Improving the security of mobile phones is one of the crucial points required to assure the personal information and the operations that can be performed from them. This article presents an authentication procedure consisting of verifying the identity of people by making a signature in the air while holding the mobile phone. Different temporal distance algorithms have been proposed and evaluated through a database of 50 people making their signatures in the air and 6 people trying to forge each of them by studying their records. Approaches based on DTW have obtained better EER results than those based on LCS (2.80% against 3.34%). Besides, different signal normalization methods have been evaluated not finding any with better EER results that when no normalization has carried out.

Leg muscle recruitment in highly trained cyclists

In this study, we examined patterns of leg muscle recruitment and co-activation, and the relationship between muscle recruitment and cadence, in highly trained cyclists. Electromyographic (EMG) activity of the tibialis anterior, tibialis posterior, peroneus longus, gastrocnemius lateralis and soleus was recorded using intramuscular electrodes, at individual preferred cadence, 57.5, 77.5 and 92.5 rev.min(-1). The influence of electrode type and location on recorded EMG was also investigated using surface and dual intramuscular recordings. Muscle recruitment patterns varied from those previously reported, but there was little variation in muscle recruitment between these highly trained cyclists. The tibialis posterior, peroneus longus and soleus were recruited in a single, short burst of activity during the downstroke. The tibialis anterior and gastrocnemius lateralis were recruited in a biphasic and alternating manner. Contrary to existing hypotheses, our results indicate little co-activation between the tibialis posterior and peroneus longus. Peak EMG amplitude increased linearly with cadence and did not decrease at individual preferred cadence. There was little variation in patterns of muscle recruitment or co-activation with changes in cadence. Intramuscular electrode location had little influence on recorded EMG. There were significant differences between surface and intramuscular recordings from the tibialis anterior and gastrocnemius lateralis, which may explain differences between our findings and those of previous studies.

Evaluating different methods of microarray data normalization

Abstract Background With the development of DNA hybridization microarray technologies, nowadays it is possible to simultaneously assess the expression levels of thousands to tens of thousands of genes. Quantitative comparison of microarrays uncovers distinct patterns of gene expression, which define different cellular phenotypes or cellular responses to drugs. Due to technical biases, normalization of the intensity levels is a pre-requisite to performing further statistical analyses. Therefore, choosing a suitable approach for normalization can be critical, deserving judicious consideration. Results Here, we considered three commonly used normalization approaches, namely: Loess, Splines and Wavelets, and two non-parametric regression methods, which have yet to be used for normalization, namely, the Kernel smoothing and Support Vector Regression. The results obtained were compared using artificial microarray data and benchmark studies. The results indicate that the Support Vector Regression is the most robust to outliers and that Kernel is the worst normalization technique, while no practical differences were observed between Loess, Splines and Wavelets. Conclusion In face of our results, the Support Vector Regression is favored for microarray normalization due to its superiority when compared to the other methods for its robustness in estimating the normalization curve.

EMG activity normalization for trunk muscles in subjects with and without back pain

Purpose: The aims of the present study were to examine electromyographic (EMG) activity of six bilateral trunk muscles during maximal contraction in three cardinal planes, and to determine the direction of contraction that gives maximal activation for each muscle. both for healthy subjects and back-pain patients. Methods: Twenty-eight healthy subjects and 15 back-pain patients performed maximum voluntary contractions in three cardinal planes, Surface EMG signals were recorded from rectus abdominis, external oblique, internal oblique, latissimus dorsi, iliocostalis lumborum, and multifidus bilaterally. Root mean square values of the EMG data were calculated to quantify I the amplitude of EMG signals. Results: For both healthy subjects and back-pain patients. one single direction of contraction was found to give the maximum EMG signals for most muscles. Rectus abdominis demonstrated maximal activity in trunk flexion, external oblique in lateral flexion. internal oblique in axial rotation, and multifidus in extension. For the latissimus dorsi and iliocostalis lumborum. maximal activity was demonstrated in more than one cardinal plane. Conclusion: This study has implications for future research involving normalization of muscle activity to maximal levels required in many trunk EMG studies. As the latissimus dorsi and iliocostalis lumborum demonstrate individual differences in the plane that gives maximal activity, these muscles may require testing in more than one plane.

18F-FDG measurement in primary lung cancer: SUV normalization to different distribution volumes

Introduction: Standard Uptake Value (SUV) is a measurement of the uptake in a tumour normalized on the basis of a distribution volume and is used to quantify 18F-Fluorodeoxiglucose (FDG) uptake in tumors, such as primary lung tumor. Several sources of error can affect its accuracy. Normalization can be based on body weight, body surface area (BSA) and lean body mass (LBM). The aim of this study is to compare the influence of 3 normalization volumes in the calculation of SUV: body weight (SUVW), BSA (SUVBSA) and LBM (SUVLBM), with and without glucose correction, in patients with known primary lung tumor. The correlation between SUV and weight, height, blood glucose level, injected activity and time between injection and image acquisition is evaluated. Methods: Sample included 30 subjects (8 female and 22 male) with primary lung tumor, with clinical indication for 18F-FDG Positron Emission Tomography (PET). Images were acquired on a Siemens Biography according to the department’s protocol. Maximum pixel SUVW was obtained for abnormal uptake focus through semiautomatic VOI with Quantification 3D isocontour (threshold 2.5). The concentration of radioactivity (kBq/ml) was obtained from SUVW, SUVBSA, SUVLBM and the glucose corrected SUV were mathematically obtained. Results: Statistically significant differences between SUVW, SUVBSA and SUVLBM and between SUVWgluc, SUVBSAgluc and SUVLBMgluc were observed (p=0.000<0.05). The blood glucose level showed significant positive correlations with SUVW (r=0.371; p=0.043) and SUVLBM (r=0.389; p=0.034). SUVBSA showed independence of variations with the blood glucose level. Conclusion: The measurement of a radiopharmaceutical tumor uptake normalized on the basis of different distribution volumes is still variable. Further investigation on this subject is recommended.

Blood pressure normalization in a large population of hypertensive patients treated with perindopril/indapamide combination: results of the OPTIMAX trial.

OBJECTIVE: To determine if the fixed-dose perindopril/indapamide combination (Per/Ind) normalizes blood pressure (BP) in the same fraction of hypertensive patients when treated in everyday practice or in controlled trials. METHODS: In this prospective trial, 17 938 hypertensive patients were treated with Per 2 mg/Ind 0.625 mg for 3-6 months. In Group 1 Per/Ind was initiated in newly diagnosed patients (n = 7032); in Group 2 Per/Ind replaced previous therapy in patients already treated but having either their BP still uncontrolled or experiencing side-effects (n = 7423); in Group 3 Per/Ind was added to previous treatment in patients with persistently high BP (n = 3483). BP was considered normalized when &lt; or = 140/90 mm Hg. A multivariate analysis for predictors of BP normalization was performed. RESULTS: Subjects were on average 62 years old and had a baseline BP of 162.3/93.6 mm Hg. After treatment with Per/Ind, BP normalization was reached in 69.6% of patients in the Initiation group, 67.5% in the Replacement Group, and 67.4% in the Add-on Group (where patients were more frequently at risk, diabetic, or with target organ damage). Mean decreases in systolic BP of 22.8 mm Hg and in diastolic BP of 12.4 mm Hg were recorded. CONCLUSIONS: This trial was established to reflect everyday clinical practice, and a treatment strategy based on the Per/Ind combination, administered as initial, replacement, or add-on therapy, led to normalization rates that were superior to those observed in Europe in routine practice. These results support recent hypertension guidelines which encourage the use of combination therapy in the management of arterial hypertension.

Efficient targeted transcript discovery via array-based normalization of RACE libraries.

Rapid amplification of cDNA ends (RACE) is a widely used approach for transcript identification. Random clone selection from the RACE mixture, however, is an ineffective sampling strategy if the dynamic range of transcript abundances is large. To improve sampling efficiency of human transcripts, we hybridized the products of the RACE reaction onto tiling arrays and used the detected exons to delineate a series of reverse-transcriptase (RT)-PCRs, through which the original RACE transcript population was segregated into simpler transcript populations. We independently cloned the products and sequenced randomly selected clones. This approach, RACEarray, is superior to direct cloning and sequencing of RACE products because it specifically targets new transcripts and often results in overall normalization of transcript abundance. We show theoretically and experimentally that this strategy leads indeed to efficient sampling of new transcripts, and we investigated multiplexing the strategy by pooling RACE reactions from multiple interrogated loci before hybridization.

Advanced feature selection methods in multinominal dementia classification from structural MRI data

Recent studies showed that features extracted from brain MRIs can well discriminate Alzheimer’s disease from Mild Cognitive Impairment. This study provides an algorithm that sequentially applies advanced feature selection methods for findings the best subset of features in terms of binary classification accuracy. The classifiers that provided the highest accuracies, have been then used for solving a multi-class problem by the one-versus-one strategy. Although several approaches based on Regions of Interest (ROIs) extraction exist, the prediction power of features has not yet investigated by comparing filter and wrapper techniques. The findings of this work suggest that (i) the IntraCranial Volume (ICV) normalization can lead to overfitting and worst the accuracy prediction of test set and (ii) the combined use of a Random Forest-based filter with a Support Vector Machines-based wrapper, improves accuracy of binary classification.