Di Wu

limma powers differential expression analyses for RNA-sequencing and microarray studies

limma is an R/Bioconductor software package that provides an integrated solution for analysing data from gene expression experiments. It contains rich features for handling complex experimental designs and for information borrowing to overcome the problem of small sample sizes. Over the past decade, limma has been a popular choice for gene discovery through differential expression analyses of microarray and high-throughput PCR data. The package contains particularly strong facilities for reading, normalizing and exploring such data. Recently, the capabilities of limma have been significantly expanded in two important directions. First, the package can now perform both differential expression and differential splicing analyses of RNA sequencing (RNA-seq) data. All the downstream analysis tools previously restricted to microarray data are now available for RNA-seq as well. These capabilities allow users to analyse both RNA-seq and microarray data with very similar pipelines. Second, the package is now able to go past the traditional gene-wise expression analyses in a variety of ways, analysing expression profiles in terms of co-regulated sets of genes or in terms of higher-order expression signatures. This provides enhanced possibilities for biological interpretation of gene expression differences. This article reviews the philosophy and design of the limma package, summarizing both new and historical features, with an emphasis on recent enhancements and features that have not been previously described.

Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers

Basal-like breast cancers arising in women carrying mutations in the BRCA1 gene, encoding the tumor suppressor protein BRCA1, are thought to develop from the mammary stem cell. To explore early cellular changes that occur in BRCA1 mutation carriers, we have prospectively isolated distinct epithelial subpopulations from normal mammary tissue and preneoplastic specimens from individuals heterozygous for a BRCA1 mutation. We describe three epithelial subsets including basal stem/progenitor, luminal progenitor and mature luminal cells. Unexpectedly, we found that breast tissue from BRCA1 mutation carriers harbors an expanded luminal progenitor population that shows factor-independent growth in vitro. Moreover, gene expression profiling revealed that breast tissue heterozygous for a BRCA1 mutation and basal breast tumors were more similar to normal luminal progenitor cells than any other subset, including the stem cell–enriched population. The c-KIT tyrosine kinase receptor (encoded by KIT) emerged as a key marker of luminal progenitor cells and was more highly expressed in BRCA1-associated preneoplastic tissue and tumors. Our findings suggest that an aberrant luminal progenitor population is a target for transformation in BRCA1-associated basal tumors .

Control of mammary stem cell function by steroid hormone signalling

The ovarian hormones oestrogen and progesterone profoundly influence breast cancer risk, underpinning the benefit of endocrine therapies in the treatment of breast cancer. Modulation of their effects through ovarian ablation or chemoprevention strategies also significantly decreases breast cancer incidence. Conversely, there is an increased risk of breast cancer associated with pregnancy in the short term. The cellular mechanisms underlying these observations, however, are poorly defined. Here we demonstrate that mouse mammary stem cells (MaSCs) are highly responsive to steroid hormone signalling, despite lacking the oestrogen and progesterone receptors. Ovariectomy markedly diminished MaSC number and outgrowth potential in vivo, whereas MaSC activity increased in mice treated with oestrogen plus progesterone. Notably, even three weeks of treatment with the aromatase inhibitor letrozole was sufficient to reduce the MaSC pool. In contrast, pregnancy led to a transient 11-fold increase in MaSC numbers, probably mediated through paracrine signalling from RANK ligand. The augmented MaSC pool indicates a cellular basis for the short-term increase in breast cancer incidence that accompanies pregnancy. These findings further indicate that breast cancer chemoprevention may be achieved, in part, through suppression of MaSC function.

Transcriptome analyses of mouse and human mammary cell subpopulations reveal multiple conserved genes and pathways

IntroductionMolecular characterization of the normal epithelial cell types that reside in the mammary gland is an important step toward understanding pathways that regulate self-renewal, lineage commitment, and differentiation along the hierarchy. Here we determined the gene expression signatures of four distinct subpopulations isolated from the mouse mammary gland. The epithelial cell signatures were used to interrogate mouse models of mammary tumorigenesis and to compare with their normal human counterpart subsets to identify conserved genes and networks.MethodsRNA was prepared from freshly sorted mouse mammary cell subpopulations (mammary stem cell (MaSC)-enriched, committed luminal progenitor, mature luminal and stromal cell) and used for gene expression profiling analysis on the Illumina platform. Gene signatures were derived and compared with those previously reported for the analogous normal human mammary cell subpopulations. The mouse and human epithelial subset signatures were then subjected to Ingenuity Pathway Analysis (IPA) to identify conserved pathways.ResultsThe four mouse mammary cell subpopulations exhibited distinct gene signatures. Comparison of these signatures with the molecular profiles of different mouse models of mammary tumorigenesis revealed that tumors arising in MMTV-Wnt-1 and p53-/- mice were enriched for MaSC-subset genes, whereas the gene profiles of MMTV-Neu and MMTV-PyMT tumors were most concordant with the luminal progenitor cell signature. Comparison of the mouse mammary epithelial cell signatures with their human counterparts revealed substantial conservation of genes, whereas IPA highlighted a number of conserved pathways in the three epithelial subsets.ConclusionsThe conservation of genes and pathways across species further validates the use of the mouse as a model to study mammary gland development and highlights pathways that are likely to govern cell-fate decisions and differentiation. It is noteworthy that many of the conserved genes in the MaSC population have been considered as epithelial-mesenchymal transition (EMT) signature genes. Therefore, the expression of these genes in tumor cells may reflect basal epithelial cell characteristics and not necessarily cells that have undergone an EMT. Comparative analyses of normal mouse epithelial subsets with murine tumor models have implicated distinct cell types in contributing to tumorigenesis in the different models.

ROAST: rotation gene set tests for complex microarray experiments

Motivation: A gene set test is a differential expression analysis in which a P-value is assigned to a set of genes as a unit. Gene set tests are valuable for increasing statistical power, organizing and interpreting results and for relating expression patterns across different experiments. Existing methods are based on permutation. Methods that rely on permutation of probes unrealistically assume independence of genes, while those that rely on permutation of sample are suitable only for two-group comparisons with a good number of replicates in each group. Results: We present ROAST, a statistically rigorous gene set test that allows for gene-wise correlation while being applicable to almost any experimental design. Instead of permutation, ROAST uses rotation, a Monte Carlo technology for multivariate regression. Since the number of rotations does not depend on sample size, ROAST gives useful results even for experiments with minimal replication. ROAST allows for any experimental design that can be expressed as a linear model, and can also incorporate array weights and correlated samples. ROAST can be tuned for situations in which only a subset of the genes in the set are actively involved in the molecular pathway. ROAST can test for uni- or bi-direction regulation. Probes can also be weighted to allow for prior importance. The power and size of the ROAST procedure is demonstrated in a simulation study, and compared to that of a representative permutation method. Finally, ROAST is used to test the degree of transcriptional conservation between human and mouse mammary stems. Availability: ROAST is implemented as a function in the Bioconductor package limma available from www.bioconductor.org Contact: smyth@wehi.edu.au Supplementary information: Supplementary data are available at Bioinformatics online.

Sequence determinants of improved CRISPR sgRNA design

The CRISPR/Cas9 system has revolutionized mammalian somatic cell genetics. Genome-wide functional screens using CRISPR/Cas9-mediated knockout or dCas9 fusion-mediated inhibition/activation (CRISPRi/a) are powerful techniques for discovering phenotype-associated gene function. We systematically assessed the DNA sequence features that contribute to single guide RNA (sgRNA) efficiency in CRISPR-based screens. Leveraging the information from multiple designs, we derived a new sequence model for predicting sgRNA efficiency in CRISPR/Cas9 knockout experiments. Our model confirmed known features and suggested new features including a preference for cytosine at the cleavage site. The model was experimentally validated for sgRNA-mediated mutation rate and protein knockout efficiency. Tested on independent data sets, the model achieved significant results in both positive and negative selection conditions and outperformed existing models. We also found that the sequence preference for CRISPRi/a is substantially different from that for CRISPR/Cas9 knockout and propose a new model for predicting sgRNA efficiency in CRISPRi/a experiments. These results facilitate the genome-wide design of improved sgRNA for both knockout and CRISPRi/a studies.

Camera: a competitive gene set test accounting for inter-gene correlation

Competitive gene set tests are commonly used in molecular pathway analysis to test for enrichment of a particular gene annotation category amongst the differential expression results from a microarray experiment. Existing gene set tests that rely on gene permutation are shown here to be extremely sensitive to inter-gene correlation. Several data sets are analyzed to show that inter-gene correlation is non-ignorable even for experiments on homogeneous cell populations using genetically identical model organisms. A new gene set test procedure (CAMERA) is proposed based on the idea of estimating the inter-gene correlation from the data, and using it to adjust the gene set test statistic. An efficient procedure is developed for estimating the inter-gene correlation and characterizing its precision. CAMERA is shown to control the type I error rate correctly regardless of inter-gene correlations, yet retains excellent power for detecting genuine differential expression. Analysis of breast cancer data shows that CAMERA recovers known relationships between tumor subtypes in very convincing terms. CAMERA can be used to analyze specified sets or as a pathway analysis tool using a database of molecular signatures.

Gata-3 Negatively Regulates the Tumor-Initiating Capacity of Mammary Luminal Progenitor Cells and Targets the Putative Tumor Suppressor Caspase-14

ABSTRACT The transcription factor Gata-3 is a definitive marker of luminal breast cancers and a key regulator of mammary morphogenesis. Here we have explored a role for Gata-3 in tumor initiation and the underlying cellular mechanisms using a mouse model of “luminal-like” cancer. Loss of a single Gata-3 allele markedly accelerated tumor progression in mice carrying the mouse mammary tumor virus promoter-driven polyomavirus middle T antigen (MMTV-PyMT mice), while overexpression of Gata-3 curtailed tumorigenesis. Through the identification of two distinct luminal progenitor cells in the mammary gland, we demonstrate that Gata-3 haplo-insufficiency increases the tumor-initiating capacity of these progenitors but not the stem cell-enriched population. Overexpression of a conditional Gata-3 transgene in the PyMT model promoted cellular differentiation and led to reduced tumor-initiating capacity as well as diminished angiogenesis. Transcript profiling studies identified caspase-14 as a novel downstream target of Gata-3, in keeping with its roles in differentiation and tumorigenesis. A strong association was evident between GATA-3 and caspase-14 expression in preinvasive ductal carcinoma in situ samples, where GATA-3 also displayed prognostic significance. Overall, these studies identify GATA-3 as an important regulator of tumor initiation through its ability to promote the differentiation of committed luminal progenitor cells.

The use of miRNA microarrays for the analysis of cancer samples with global miRNA decrease

Recent studies have established that mutations or deletions in microRNA (miRNA) processing enzymes resulting in a global decrease of miRNA expression are frequent across cancers and can be associated with a poorer prognosis. While very popular in miRNA profiling studies, it remains unclear whether miRNA microarrays are suited or not to accurately detecting global miRNA decreases seen in cancers. In this work, we analyzed the miRNA profiles of samples with global miRNA decreases using Affymetrix miRNA microarrays following the inducible genetic deletion of Dicer1. Surprisingly, up to a third of deregulated miRNAs identified upon Dicer1 depletion were found to be up-regulated following standard robust multichip average (RMA) background correction and quantile normalization, indicative of normalization bias. Our comparisons of five preprocess steps performed at the probe level demonstrated that the use of cyclic loess relying on non-miRNA small RNAs present on the Affymetrix platform significantly improved specificity and sensitivity of detection of decreased miRNAs. These findings were validated in samples from patients with prostate cancer, where conjugation of robust normal-exponential background correction with cyclic loess normalization and array weights correctly identified the greatest number of decreased miRNAs, and the lowest amount of false-positive up-regulated miRNAs. These findings highlight the importance of miRNA microarray normalization for the detection of miRNAs that are truly differentially expressed and suggest that the use of cyclic loess based on non-miRNA small RNAs can help to improve the sensitivity and specificity of miRNA profiling in cancer samples with global miRNA decrease.

Statistical inference for time course RNA-Seq data using a negative binomial mixed-effect model

BackgroundAccurate identification of differentially expressed (DE) genes in time course RNA-Seq data is crucial for understanding the dynamics of transcriptional regulatory network. However, most of the available methods treat gene expressions at different time points as replicates and test the significance of the mean expression difference between treatments or conditions irrespective of time. They thus fail to identify many DE genes with different profiles across time. In this article, we propose a negative binomial mixed-effect model (NBMM) to identify DE genes in time course RNA-Seq data. In the NBMM, mean gene expression is characterized by a fixed effect, and time dependency is described by random effects. The NBMM is very flexible and can be fitted to both unreplicated and replicated time course RNA-Seq data via a penalized likelihood method. By comparing gene expression profiles over time, we further classify the DE genes into two subtypes to enhance the understanding of expression dynamics. A significance test for detecting DE genes is derived using a Kullback-Leibler distance ratio. Additionally, a significance test for gene sets is developed using a gene set score.ResultsSimulation analysis shows that the NBMM outperforms currently available methods for detecting DE genes and gene sets. Moreover, our real data analysis of fruit fly developmental time course RNA-Seq data demonstrates the NBMM identifies biologically relevant genes which are well justified by gene ontology analysis.ConclusionsThe proposed method is powerful and efficient to detect biologically relevant DE genes and gene sets in time course RNA-Seq data.