Calibration of microarray gene-expression data

Variation of RNA Quality and Quantity Are Major Sources of Batch Effects in Microarray Expression Data

The great utility of microarrays for genome-scale expression analysis is challenged by the widespread presence of batch effects, which bias expression measurements in particular within large data sets. These unwanted technical artifacts can obscure biological variation and thus significantly reduce the reliability of the analysis results. It is largely unknown which are the predominant technical sources leading to batch effects. We here quantitatively assess the prevalence and impact of several known technical effects on microarray expression results. Particularly, we focus on important factors such as RNA degradation, RNA quantity, and sequence biases including multiple guanine effects. We find that the common variation of RNA quality and RNA quantity can not only yield low-quality expression results, but that both factors also correlate with batch effects and biological characteristics of the samples.

The relationship between glial cell mechanosensitivity and foreign body reactions in the central nervous system

Devices implanted into the body become encapsulated due to a foreign body reaction. In the central nervous system (CNS), this can lead to loss of functionality in electrodes used to treat disorders. Around CNS implants, glial cells are activated, undergo gliosis and ultimately encapsulate the electrodes. The primary cause of this reaction is unknown. Here we show that the mechanical mismatch between nervous tissue and electrodes activates glial cells. Both primary rat microglial cells and astrocytes responded to increasing the contact stiffness from physiological values (G' ∼ 100 Pa) to shear moduli G' ≥ 10 kPa by changes in morphology and upregulation of inflammatory genes and proteins. Upon implantation of composite foreign bodies into rat brains, foreign body reactions were significantly enhanced around their stiff portions in vivo. Our results indicate that CNS glial cells respond to mechanical cues, and suggest that adapting the surface stiffness of neural implants to that of nervous tissue could minimize adverse reactions and improve biocompatibility.

Analysis of microRNA expression using machine learning

The systematic analysis of miRNA expression and its potential mRNA targets constitutes a basal objective in miRNA research in addition to miRNA gene detection and miRNA target prediction. In this chapter we address methodical issues of miRNA expression analysis using self-organizing maps (SOM), a neural network machine learning algorithm with strong visualization and second-level analysis capabilities widely used to categorize large-scale, high-dimensional data. We shortly review selected experimental and theoretical aspects of miRNA expression analysis. Then, the protocol of our SOM method is outlined with special emphasis on miRNA/mRNA coexpression. The method allows extracting differentially expressed RNA transcripts, their functional context, and also characterization of global properties of expression states and profiles. In addition to the separate study of miRNA and mRNA expression landscapes, we propose the combined analysis of both entities using a covariance SOM.

Identification of micro-RNA networks in end-stage heart failure because of dilated cardiomyopathy

Micro-RNAs regulate gene expression by directly binding to the target mRNAs. The goal of the study was to examine the expression profiling of miRNAs in human failing hearts and identify the key miRNAs that regulate molecular signalling networks and thus contribute to this pathological process. The levels of miRNAs and expressed genes were analysed in myocardial biopsy samples from patients with end-stage heart failure (n = 14) and those from normal heart samples (n = 8). Four networks were built including the Gene regulatory network, Signal-Network, miRNA-GO-Network and miRNA-Gene-Network. According to the fold change in the network and probability values in the microarray cohort, RT-PCR was performed to measure the expression of five of the 72 differentially regulated miRNAs. miR-340 achieved statistically significant. miR-340 was identified for the first time in cardiac pathophysiological condition. We overexpressed miR-340 in cultured neonatal rat cardiomyocytes to identify whether miR-340 plays a determining role in the progression of heart failure. ANP, BNP and caspase-3 were significantly elevated in the miR-340 transfected cells compared with controls (P < 0.05). The cross-sectional area of overexpressing miR-340 cardiomyocytes (1952.22 ± 106.59) was greater (P < 0.0001) than controls (1059.99 ± 45.59) documented by Laser Confocal Microscopy. The changes of cellular structure and the volume were statistical significance. Our study provided a comprehensive miRNA expression profiling in the end-stage heart failure and identified miR-340 as a key miRNA contributing to the occurrence and progression of heart failure. Our discoveries provide novel therapeutic targets for patients with heart failure.

Isolation and analysis of mRNAs from specific cell types of plants by ribosome immunopurification

Multiple ribosomes assemble onto an individual mRNA to form a polyribosome (polysome) complex. The epitope tagging of specific ribosomal proteins can enable the immunopurification of polysomes from crude cell extracts derived from cryopreserved tissue samples. Through expression of the epitope-tagged ribosomal protein in cell-type and regional specific domains of Arabidopsis thaliana and other organisms it is feasible to quantitatively assess the mRNAs that are associated with ribosomes with cell-specific resolution. Here we present detailed methods for development of transgenics that express a FLAG-tagged version of ribosomal protein L18 (RPL18) under the direction of individual promoters with specific domains of expression, the immunopurification of ribosomes, and bioinformatic analyses of the resultant datasets obtained by microarray profiling. This methodology provides researchers with the opportunity to assess rapid changes at the organ, tissue, regional or cell-type specific level of mRNAs that are associated with ribosomes and therefore engaged in translation.

Mining SOM expression portraits: feature selection and integrating concepts of molecular function

Background

Self organizing maps (SOM) enable the straightforward portraying of high-dimensional data of large sample collections in terms of sample-specific images. The analysis of their texture provides so-called spot-clusters of co-expressed genes which require subsequent significance filtering and functional interpretation. We address feature selection in terms of the gene ranking problem and the interpretation of the obtained spot-related lists using concepts of molecular function.

Results

Different expression scores based either on simple fold change-measures or on regularized Student’s t-statistics are applied to spot-related gene lists and compared with special emphasis on the error characteristics of microarray expression data. The spot-clusters are analyzed using different methods of gene set enrichment analysis with the focus on overexpression and/or overrepresentation of predefined sets of genes. Metagene-related overrepresentation of selected gene sets was mapped into the SOM images to assign gene function to different regions. Alternatively we estimated set-related overexpression profiles over all samples studied using a gene set enrichment score. It was also applied to the spot-clusters to generate lists of enriched gene sets. We used the tissue body index data set, a collection of expression data of human tissues as an illustrative example. We found that tissue related spots typically contain enriched populations of gene sets well corresponding to molecular processes in the respective tissues. In addition, we display special sets of housekeeping and of consistently weak and high expressed genes using SOM data filtering.

Conclusions

The presented methods allow the comprehensive downstream analysis of SOM-transformed expression data in terms of cluster-related gene lists and enriched gene sets for functional interpretation. SOM clustering implies the ability to define either new gene sets using selected SOM spots or to verify and/or to amend existing ones.

Stable feature selection and classification algorithms for multiclass microarray data

Background

Recent studies suggest that gene expression profiles are a promising alternative for clinical cancer classification. One major problem in applying DNA microarrays for classification is the dimension of obtained data sets. In this paper we propose a multiclass gene selection method based on Partial Least Squares (PLS) for selecting genes for classification. The new idea is to solve multiclass selection problem with the PLS method and decomposition to a set of two-class sub-problems: one versus rest (OvR) and one versus one (OvO). We use OvR and OvO two-class decomposition for other recently published gene selection method. Ranked gene lists are highly unstable in the sense that a small change of the data set often leads to big changes in the obtained ordered lists. In this paper, we take a look at the assessment of stability of the proposed methods. We use the linear support vector machines (SVM) technique in different variants: one versus one, one versus rest, multiclass SVM (MSVM) and the linear discriminant analysis (LDA) as a classifier. We use balanced bootstrap to estimate the prediction error and to test the variability of the obtained ordered lists.

Results

This paper focuses on effective identification of informative genes. As a result, a new strategy to find a small subset of significant genes is designed. Our results on real multiclass cancer data show that our method has a very high accuracy rate for different combinations of classification methods, giving concurrently very stable feature rankings.

Conclusions

This paper shows that the proposed strategies can improve the performance of selected gene sets substantially. OvR and OvO techniques applied to existing gene selection methods improve results as well. The presented method allows to obtain a more reliable classifier with less classifier error. In the same time the method generates more stable ordered feature lists in comparison with existing methods.

Reviewers

This article was reviewed by Prof Marek Kimmel, Dr Hans Binder (nominated by Dr Tomasz Lipniacki) and Dr Yuriy Gusev

The testis-specific double bromodomain-containing protein BRDT forms a complex with multiple spliceosome components and is required for mRNA splicing and 3'-UTR truncation in round spermatids

Members of the BET (bromodomain and extra terminal motif) family of proteins have been shown to be chromatin-interacting regulators of transcription. We previously generated a mutation in the testis-specific mammalian BET gene Brdt (bromodomain, testis-specific) that yields protein lacking the first bromodomain (BRDT(ΔBD1)) and observed disrupted spermiogenesis and male sterility. To determine whether BRDT(ΔBD1) protein results in altered transcription, we analyzed the transcriptomes of control versus Brdt(ΔBD1/ΔBD1) round spermatids. Over 400 genes showed statistically significant differential expression, and among the up-regulated genes, there was an enrichment of RNA splicing genes. Over 60% of these splicing genes had transcripts that lacked truncation of their 3'-untranslated region (UTR) typical of round spermatids. We selected four of these genes to characterize: Srsf2, Ddx5, Hnrnpk and Tardbp. The 3'-UTRs of Srsf2, Ddx5 and Hnrnpk mRNAs were longer in mutant round spermatids and resulted in reduced protein levels. Tardbp was transcriptionally up-regulated and a splicing shift toward the longer variant was observed. All four splicing proteins were found to complex with BRDT in control and mutant testes. We thus suggest that, along with modulating transcription, BRDT modulates gene expression as part of the splicing machinery. These modulations alter 3'-UTR processing in round spermatids; importantly, the BD1 is essential for these functions.

[Translational research in head and neck cancer. Biological characteristics and general aspects]

Translational research in head and neck oncology is subject to the same laws as all other solid tumors. It is based on the one hand on a solid framework of well prepared clinical studies and / or workflows according to consensus criteria with comparable documentation of clinical outcomes, while on the other on methodolgically solid and reproducible laboratory research within an effeciently interacting network. Translationally applicable single molecular markers from basic research [with the exception of p16(INK4a) as a surrogate marker for human papillomavirus (HPV)] have not found their way into clinical routine in head and neck squamous cell carcinoma (HNSCC). "Correlated gene sets" and "metagenes", including genetic profiling (omics) within clinically characterized patient groups, play an increasing role in the translational research of HNSCC. Although methodological problems currently hinder clinical oncological research, increasing focus on translational research can be observed.

Transcriptomics using next generation sequencing technologies

Next generation sequencing technologies may now be applied to the study of transcriptomics. RNA-Seq or RNA sequencing employs high-throughput sequencing of complementary DNA fragments delivering a transcriptional profile. In this chapter, we aim to provide a starting point for Xenopus researchers planning on starting an RNA-Seq transcriptomics study. We begin by providing a section on template isolation and library preparation. The next section comprises the main bioinformatics procedures that need to be performed for raw data processing, normalization, and differential gene expression. Finally, we have included a section on studying deep sequencing results in Xenopus, which offers general guidance as to what can be done in this model.

Analysis of metabolic and gene expression changes after hydrodynamic DNA injection into mouse liver

The hydrodynamic injection in mice tail vein of a plasmid (40 µg DNA) bearing the human α1-antitrypsin gene mediates: a) good liver gene transfer resulting in therapeutic plasma levels of human protein (1 mg/ml, approximately) from days 1-10 after injection; b) low liver injury as demonstrated by a poor and transient increase of aspartate aminotransferase (AST) and alanine transaminase (ALT) in mouse plasma; 3) limited expression and metabolic changes in host liver genes and metabolites as evaluated on days 2 and 10 after injection. Groups of three mice were uninjected (control) or hydrodynamically injected with saline or plasmid DNA and then sacrificed on days 2 and 10 after injection. The results of principal component analysis (PCA) show, both in expression microarray and metabolomic analysis, that changes between control and hydrodynamically injected groups are not dramatic and tend to normalize after 10 d. The differences are even smaller between DNA and saline hydrodynamically injected mice. Hydrodynamic injection induces a complex but limited gene expression and metabolic change which includes variations in molecules related to energy metabolism and stress response. The results contribute to support that hydrodynamic method is a safe procedure of liver gene transfer but the long-term effect of hydrodynamic gene transfer procedure, remains to be studied.

Experimental design, preprocessing, normalization and differential expression analysis of small RNA sequencing experiments

Prior to the advent of new, deep sequencing methods, small RNA (sRNA) discovery was dependent on Sanger sequencing, which was time-consuming and limited knowledge to only the most abundant sRNA. The innovation of large-scale, next-generation sequencing has exponentially increased knowledge of the biology, diversity and abundance of sRNA populations. In this review, we discuss issues involved in the design of sRNA sequencing experiments, including choosing a sequencing platform, inherent biases that affect sRNA measurements and replication. We outline the steps involved in preprocessing sRNA sequencing data and review both the principles behind and the current options for normalization. Finally, we discuss differential expression analysis in the absence and presence of biological replicates. While our focus is on sRNA sequencing experiments, many of the principles discussed are applicable to the sequencing of other RNA populations.

G-stack modulated probe intensities on expression arrays - sequence corrections and signal calibration

Background

The brightness of the probe spots on expression microarrays intends to measure the abundance of specific mRNA targets. Probes with runs of at least three guanines (G) in their sequence show abnormal high intensities which reflect rather probe effects than target concentrations. This G-bias requires correction prior to downstream expression analysis.

Results

Longer runs of three or more consecutive G along the probe sequence and in particular triple degenerated G at its solution end ((GGG)₁-effect) are associated with exceptionally large probe intensities on GeneChip expression arrays. This intensity bias is related to non-specific hybridization and affects both perfect match and mismatch probes. The (GGG)₁-effect tends to increase gradually for microarrays of later GeneChip generations. It was found for DNA/RNA as well as for DNA/DNA probe/target-hybridization chemistries. Amplification of sample RNA using T7-primers is associated with strong positive amplitudes of the G-bias whereas alternative amplification protocols using random primers give rise to much smaller and partly even negative amplitudes.

We applied positional dependent sensitivity models to analyze the specifics of probe intensities in the context of all possible short sequence motifs of one to four adjacent nucleotides along the 25meric probe sequence. Most of the longer motifs are adequately described using a nearest-neighbor (NN) model. In contrast, runs of degenerated guanines require explicit consideration of next nearest neighbors (GGG terms). Preprocessing methods such as vsn, RMA, dChip, MAS5 and gcRMA only insufficiently remove the G-bias from data.

Conclusions

Positional and motif dependent sensitivity models accounts for sequence effects of oligonucleotide probe intensities. We propose a positional dependent NN+GGG hybrid model to correct the intensity bias associated with probes containing poly-G motifs. It is implemented as a single-chip based calibration algorithm for GeneChips which can be applied in a pre-correction step prior to standard preprocessing.

"Hook"-calibration of GeneChip-microarrays: chip characteristics and expression measures

Background

Microarray experiments rely on several critical steps that may introduce biases and uncertainty in downstream analyses. These steps include mRNA sample extraction, amplification and labelling, hybridization, and scanning causing chip-specific systematic variations on the raw intensity level. Also the chosen array-type and the up-to-dateness of the genomic information probed on the chip affect the quality of the expression measures. In the accompanying publication we presented theory and algorithm of the so-called hook method which aims at correcting expression data for systematic biases using a series of new chip characteristics.

Results

In this publication we summarize the essential chip characteristics provided by this method, analyze special benchmark experiments to estimate transcript related expression measures and illustrate the potency of the method to detect and to quantify the quality of a particular hybridization. It is shown that our single-chip approach provides expression measures responding linearly on changes of the transcript concentration over three orders of magnitude. In addition, the method calculates a detection call judging the relation between the signal and the detection limit of the particular measurement. The performance of the method in the context of different chip generations and probe set assignments is illustrated. The hook method characterizes the RNA-quality in terms of the 3'/5'-amplification bias and the sample-specific calling rate. We show that the proper judgement of these effects requires the disentanglement of non-specific and specific hybridization which, otherwise, can lead to misinterpretations of expression changes. The consequences of modifying probe/target interactions by either changing the labelling protocol or by substituting RNA by DNA targets are demonstrated.

Conclusion

The single-chip based hook-method provides accurate expression estimates and chip-summary characteristics using the natural metrics given by the hybridization reaction with the potency to develop new standards for microarray quality control and calibration.