A flexible approach for predictive biomarker discovery.

An endeavor central to precision medicine is predictive biomarker discovery; they define patient subpopulations which stand to benefit most, or least, from a given treatment. The identification of these biomarkers is often the byproduct of the related but fundamentally different task of treatment rule estimation. Using treatment rule estimation methods to identify predictive biomarkers in clinical trials where the number of covariates exceeds the number of participants often results in high false discovery rates. The higher than expected number of false positives translates to wasted resources when conducting follow-up experiments for drug target identification and diagnostic assay development. Patient outcomes are in turn negatively affected. We propose a variable importance parameter for directly assessing the importance of potentially predictive biomarkers and develop a flexible nonparametric inference procedure for this estimand. We prove that our estimator is double robust and asymptotically linear under loose conditions in the data-generating process, permitting valid inference about the importance metric. The statistical guarantees of the method are verified in a thorough simulation study representative of randomized control trials with moderate and high-dimensional covariate vectors. Our procedure is then used to discover predictive biomarkers from among the tumor gene expression data of metastatic renal cell carcinoma patients enrolled in recently completed clinical trials. We find that our approach more readily discerns predictive from nonpredictive biomarkers than procedures whose primary purpose is treatment rule estimation. An open-source software implementation of the methodology, the uniCATE R package, is briefly introduced.

Practical guidance on modeling choices for the virtual twins method.

Individuals can vary drastically in their response to the same treatment, and this heterogeneity has driven the push for more personalized medicine. Accurate and interpretable methods to identify subgroups that respond to the treatment differently from the population average are necessary to achieving this goal. The Virtual Twins (VT) method is a highly cited and implemented method for subgroup identification because of its intuitive framework. However, since its initial publication, many researchers still rely heavily on the authors' initial modeling suggestions without examining newer and more powerful alternatives. This leaves much of the potential of the method untapped. We comprehensively evaluate the performance of VT with different combinations of methods in each of its component steps, under a collection of linear and nonlinear problem settings. Our simulations show that the method choice for Step 1 of VT, in which dense models with high predictive performance are fit for the potential outcomes, is highly influential in the overall accuracy of the method, and Superlearner is a promising choice. We illustrate our findings by using VT to identify subgroups with heterogeneous treatment effects in a randomized, double-blind trial of very low nicotine content cigarettes.

Time-resolved photoconductivity distribution measurement by a synchronized double-scanning method.

The lifetime and the distribution of photoconductivity generated in laser-illuminated semiconductors are critical to photoconductivity-based applications. We propose a synchronized double-scanning method to measure time-resolved diffusion in the form of an afterglow embedded in the distribution map. The method combines spatial scanning of a coaxial resonator with synchronized laser scanning to map the dynamically excited conductivity on a semiconductor wafer. Thus, the photoconductivity afterglow effects can be mapped and retrieved by images of dynamic photoconductivity distribution. The photoconductivity lifetimes of silicon wafers with different thicknesses and by different lasers were measured and evaluated, which were also validated by the microwave photoconductivity decay (µ-PCD) method. In addition, the behavior of photoconductivity diffusion around a structural defect was exhibited. The method is nondestructive and can be applied in the photoconductivity property diagnostic.

Real-time imaging of electromagnetic fields.

The measurement and diagnosis of electromagnetic fields are important foundations for various electronic and optical systems. This paper presents an innovative optically controlled plasma scattering technique for imaging electromagnetic fields. On a silicon wafer, the plasma induced by the photoconductive effect is exploited as an optically controlled scattering probe to image the amplitude and phase of electromagnetic fields. A prototype is built and realizes the imaging of electromagnetic fields radiated from antennas from 870MHz to 0.2 terahertz within one second. Measured results show good agreement with the simulations. It is demonstrated that this new technology improves the efficiency of electromagnetic imaging to a real-time level, while combining various advantages of ultrafast speed, super-resolution, ultra-wideband response, low-cost and vectorial wave mapping ability. This method may initiate a new avenue in the measurement and diagnosis of electromagnetic fields.

SCnorm: robust normalization of single-cell RNA-seq data.

The normalization of RNA-seq data is essential for accurate downstream inference, but the assumptions upon which most normalization methods are based are not applicable in the single-cell setting. Consequently, applying existing normalization methods to single-cell RNA-seq data introduces artifacts that bias downstream analyses. To address this, we introduce SCnorm for accurate and efficient normalization of single-cell RNA-seq data.

Automated minute scale RNA-seq of pluripotent stem cell differentiation reveals early divergence of human and mouse gene expression kinetics.

Pluripotent stem cells retain the developmental timing of their species of origin in vitro, an observation that suggests the existence of a cell-intrinsic developmental clock, yet the nature and machinery of the clock remain a mystery. We hypothesize that one possible component may lie in species-specific differences in the kinetics of transcriptional responses to differentiation signals. Using a liquid-handling robot, mouse and human pluripotent stem cells were exposed to identical neural differentiation conditions and sampled for RNA-sequencing at high frequency, every 4 or 10 minutes, for the first 10 hours of differentiation to test for differences in transcriptomic response rates. The majority of initial transcriptional responses occurred within a rapid window in the first minutes of differentiation for both human and mouse stem cells. Despite similarly early onsets of gene expression changes, we observed shortened and condensed gene expression patterns in mouse pluripotent stem cells compared to protracted trends in human pluripotent stem cells. Moreover, the speed at which individual genes were upregulated, as measured by the slopes of gene expression changes over time, was significantly faster in mouse compared to human cells. These results suggest that downstream transcriptomic response kinetics to signaling cues are faster in mouse versus human cells, and may offer a partial account for the vast differences in developmental rates across species.

Functional characterization of human pluripotent stem cell-derived arterial endothelial cells.

Here, we report the derivation of arterial endothelial cells from human pluripotent stem cells that exhibit arterial-specific functions in vitro and in vivo. We combine single-cell RNA sequencing of embryonic mouse endothelial cells with an EFNB2-tdTomato/EPHB4-EGFP dual reporter human embryonic stem cell line to identify factors that regulate arterial endothelial cell specification. The resulting xeno-free protocol produces cells with gene expression profiles, oxygen consumption rates, nitric oxide production levels, shear stress responses, and TNFα-induced leukocyte adhesion rates characteristic of arterial endothelial cells. Arterial endothelial cells were robustly generated from multiple human embryonic and induced pluripotent stem cell lines and have potential applications for both disease modeling and regenerative medicine.

Atezolizumab versus chemotherapy in patients with platinum-treated locally advanced or metastatic urothelial carcinoma (IMvigor211): a multicentre, open-label, phase 3 randomised controlled trial.

BACKGROUND: Few options exist for patients with locally advanced or metastatic urothelial carcinoma after progression with platinum-based chemotherapy. We aimed to assess the safety and efficacy of atezolizumab (anti-programmed death-ligand 1 [PD-L1]) versus chemotherapy in this patient population. METHODS: We conducted this multicentre, open-label, phase 3 randomised controlled trial (IMvigor211) at 217 academic medical centres and community oncology practices mainly in Europe, North America, and the Asia-Pacific region. Patients (aged ≥18 years) with metastatic urothelial carcinoma who had progressed after platinum-based chemotherapy were randomly assigned (1:1), via an interactive voice and web response system with a permuted block design (block size of four), to receive atezolizumab 1200 mg or chemotherapy (physician's choice: vinflunine 320 mg/m2, paclitaxel 175 mg/m2, or 75 mg/m2 docetaxel) intravenously every 3 weeks. Randomisation was stratified by PD-L1 expression (expression on <1% [IC0] or 1% to <5% [IC1] of tumour-infiltrating immune cells vs ≥5% of tumour-infiltrating immune cells [IC2/3]), chemotherapy type (vinflunine vs taxanes), liver metastases (yes vs no), and number of prognostic factors (none vs one, two, or three). Patients and investigators were aware of group allocation. Patients, investigators, and the sponsor were masked to PD-L1 expression status. The primary endpoint of overall survival was tested hierarchically in prespecified populations: IC2/3, followed by IC1/2/3, followed by the intention-to-treat population. This study, which is ongoing but not recruiting participants, is registered with ClinicalTrials.gov, number NCT02302807. FINDINGS: Between Jan 13, 2015, and Feb 15, 2016, we randomly assigned 931 patients from 198 sites to receive atezolizumab (n=467) or chemotherapy (n=464). In the IC2/3 population (n=234), overall survival did not differ significantly between patients in the atezolizumab group and those in the chemotherapy group (median 11·1 months [95% CI 8·6-15·5; n=116] vs 10·6 months [8·4-12·2; n=118]; stratified hazard ratio [HR] 0·87, 95% CI 0·63-1·21; p=0·41), thus precluding further formal statistical analysis. Confirmed objective response rates were similar between treatment groups in the IC2/3 population: 26 (23%) of 113 evaluable patients had an objective response in the atezolizumab group compared with 25 (22%) of 116 patients in the chemotherapy group. Duration of response was numerically longer in the atezolizumab group than in the chemotherapy group (median 15·9 months [95% CI 10·4 to not estimable] vs 8·3 months [5·6-13·2]; HR 0·57, 95% CI 0·26-1·26). In the intention-to-treat population, patients receiving atezolizumab had fewer grade 3-4 treatment-related adverse events than did those receiving chemotherapy (91 [20%] of 459 vs 189 [43%] of 443 patients), and fewer adverse events leading to treatment discontinuation (34 [7%] vs 78 [18%] patients). INTERPRETATION: Atezolizumab was not associated with significantly longer overall survival than chemotherapy in patients with platinum-refractory metastatic urothelial carcinoma overexpressing PD-L1 (IC2/3). However, the safety profile for atezolizumab was favourable compared with chemotherapy, Exploratory analysis of the intention-to-treat population showed well-tolerated, durable responses in line with previous phase 2 data for atezolizumab in this setting. FUNDING: F Hoffmann-La Roche, Genentech.

The Transcription Factor Nfatc2 Regulates ß-Cell Proliferation and Genes Associated with Type 2 Diabetes in Mouse and Human Islets.

Human genome-wide association studies (GWAS) have shown that genetic variation at >130 gene loci is associated with type 2 diabetes (T2D). We asked if the expression of the candidate T2D-associated genes within these loci is regulated by a common locus in pancreatic islets. Using an obese F2 mouse intercross segregating for T2D, we show that the expression of ~40% of the T2D-associated genes is linked to a broad region on mouse chromosome (Chr) 2. As all but 9 of these genes are not physically located on Chr 2, linkage to Chr 2 suggests a genomic factor(s) located on Chr 2 regulates their expression in trans. The transcription factor Nfatc2 is physically located on Chr 2 and its expression demonstrates cis linkage; i.e., its expression maps to itself. When conditioned on the expression of Nfatc2, linkage for the T2D-associated genes was greatly diminished, supporting Nfatc2 as a driver of their expression. Plasma insulin also showed linkage to the same broad region on Chr 2. Overexpression of a constitutively active (ca) form of Nfatc2 induced ß-cell proliferation in mouse and human islets, and transcriptionally regulated more than half of the T2D-associated genes. Overexpression of either ca-Nfatc2 or ca-Nfatc1 in mouse islets enhanced insulin secretion, whereas only ca-Nfatc2 was able to promote ß-cell proliferation, suggesting distinct molecular pathways mediating insulin secretion vs. ß-cell proliferation are regulated by NFAT. Our results suggest that many of the T2D-associated genes are downstream transcriptional targets of NFAT, and may act coordinately in a pathway through which NFAT regulates ß-cell proliferation in both mouse and human islets.

Overexpression of interleukin-20 receptor subunit beta (IL20RB) correlates with cell proliferation, invasion and migration enhancement and poor prognosis in papillary renal cell carcinoma.

Papillary renal cell carcinoma (PRCC) accounts for about 10 percent of all renal cell carcinomas, and the prognosis is poor for people with advanced disease. Interleukin-20 receptor subunit beta (IL20RB) is a single-pass type I membrane protein of the type II cytokine receptor family and is related to the pathogenesis of chronic inflammation and autoimmune diseases, including psoriasis, glaucoma, vitiligo, rheumatoid arthritis, and inflammatory bowel disease. However, little has been reported on IL20RB with respect to cancer, especially in PRCC. Thus, we performed this study to explore its biological characteristics in PRCC. Data from the TCGA database were used to analyze the expression and prognosis of IL20RB. qRT-PCR was used to detect the expression of IL20RB in PRCC cells in vitro. After knockdown of IL20RB with small interfering RNA (siRNA) technology, the proliferation, migration, and invasion of Ketr-3 cells and the expression of related proteins in the epithelial-mesenchymal transition (EMT) pathway were measured with Cell Counting Kit-8 (CCK-8), transwell, and western blot assays. The findings demonstrated that the expression of IL20RB was upregulated in both PRCC tissues and cells and that the high expression of IL20RB led to low overall survival (OS). Furthermore, after knockdown of IL20RB in vitro, the proliferation, migration, and invasion of Ketr-3 cells were reduced, and the expression of related proteins in the EMT pathway declined, suggesting that IL20RB plays a vital role in PRCC through the EMT pathway. These results reveal the biological significance of IL20RB in PRCC and provide new insight for future targeted drugs.

Trendy: segmented regression analysis of expression dynamics in high-throughput ordered profiling experiments.

BACKGROUND: High-throughput expression profiling experiments with ordered conditions (e.g. time-course or spatial-course) are becoming more common for studying detailed differentiation processes or spatial patterns. Identifying dynamic changes at both the individual gene and whole transcriptome level can provide important insights about genes, pathways, and critical time points. RESULTS: We present an R package, Trendy, which utilizes segmented regression models to simultaneously characterize each gene's expression pattern and summarize overall dynamic activity in ordered condition experiments. For each gene, Trendy finds the optimal segmented regression model and provides the location and direction of dynamic changes in expression. We demonstrate the utility of Trendy to provide biologically relevant results on both microarray and RNA-sequencing (RNA-seq) datasets. CONCLUSIONS: Trendy is a flexible R package which characterizes gene-specific expression patterns and summarizes changes of global dynamics over ordered conditions. Trendy is freely available on Bioconductor with a full vignette at https://bioconductor.org/packages/release/bioc/html/Trendy.html .

Analysis of embryonic development in the unsequenced axolotl: Waves of transcriptomic upheaval and stability.

The axolotl (Ambystoma mexicanum) has long been the subject of biological research, primarily owing to its outstanding regenerative capabilities. However, the gene expression programs governing its embryonic development are particularly underexplored, especially when compared to other amphibian model species. Therefore, we performed whole transcriptome polyA+ RNA sequencing experiments on 17 stages of embryonic development. As the axolotl genome is unsequenced and its gene annotation is incomplete, we built de novo transcriptome assemblies for each stage and garnered functional annotation by comparing expressed contigs with known genes in other organisms. In evaluating the number of differentially expressed genes over time, we identify three waves of substantial transcriptome upheaval each followed by a period of relative transcriptome stability. The first wave of upheaval is between the one and two cell stage. We show that the number of differentially expressed genes per unit time is higher between the one and two cell stage than it is across the mid-blastula transition (MBT), the period of zygotic genome activation. We use total RNA sequencing to demonstrate that the vast majority of genes with increasing polyA+ signal between the one and two cell stage result from polyadenylation rather than de novo transcription. The first stable phase begins after the two cell stage and continues until the mid-blastula transition, corresponding with the pre-MBT phase of transcriptional quiescence in amphibian development. Following this is a peak of differential gene expression corresponding with the activation of the zygotic genome and a phase of transcriptomic stability from stages 9-11. We observe a third wave of transcriptomic change between stages 11 and 14, followed by a final stable period. The last two stable phases have not been documented in amphibians previously and correspond to times of major morphogenic change in the axolotl embryo: gastrulation and neurulation. These results yield new insights into global gene expression during early stages of amphibian embryogenesis and will help to further develop the axolotl as a model species for developmental and regenerative biology.

Oscope identifies oscillatory genes in unsynchronized single-cell RNA-seq experiments.

Oscillatory gene expression is fundamental to development, but technologies for monitoring expression oscillations are limited. We have developed a statistical approach called Oscope to identify and characterize the transcriptional dynamics of oscillating genes in single-cell RNA-seq data from an unsynchronized cell population. Applying Oscope to a number of data sets, we demonstrated its utility and also identified a potential artifact in the Fluidigm C1 platform.

OEFinder: a user interface to identify and visualize ordering effects in single-cell RNA-seq data.

UNLABELLED: A recent article identified an artifact in multiple single-cell RNA-seq (scRNA-seq) datasets generated by the Fluidigm C1 platform. Specifically, Leng et al. showed significantly increased gene expression in cells captured from sites with small or large plate output IDs. We refer to this artifact as an ordering effect (OE). Including OE genes in downstream analyses could lead to biased results. To address this problem, we developed a statistical method and software called OEFinder to identify a sorted list of OE genes. OEFinder is available as an R package along with user-friendly graphical interface implementations which allows users to check for potential artifacts in scRNA-seq data generated by the Fluidigm C1 platform. AVAILABILITY AND IMPLEMENTATION: OEFinder is freely available at https://github.com/lengning/OEFinder CONTACT: rstewart@morgridge.org or lengning1@gmail.com SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

EBSeq-HMM: a Bayesian approach for identifying gene-expression changes in ordered RNA-seq experiments.

MOTIVATION: With improvements in next-generation sequencing technologies and reductions in price, ordered RNA-seq experiments are becoming common. Of primary interest in these experiments is identifying genes that are changing over time or space, for example, and then characterizing the specific expression changes. A number of robust statistical methods are available to identify genes showing differential expression among multiple conditions, but most assume conditions are exchangeable and thereby sacrifice power and precision when applied to ordered data. RESULTS: We propose an empirical Bayes mixture modeling approach called EBSeq-HMM. In EBSeq-HMM, an auto-regressive hidden Markov model is implemented to accommodate dependence in gene expression across ordered conditions. As demonstrated in simulation and case studies, the output proves useful in identifying differentially expressed genes and in specifying gene-specific expression paths. EBSeq-HMM may also be used for inference regarding isoform expression. AVAILABILITY AND IMPLEMENTATION: An R package containing examples and sample datasets is available at Bioconductor. CONTACT: kendzior@biostat.wisc.edu SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The gene desert mammary carcinoma susceptibility locus Mcs1a regulates Nr2f1 modifying mammary epithelial cell differentiation and proliferation.

Genome-wide association studies have revealed that many low-penetrance breast cancer susceptibility loci are located in non-protein coding genomic regions; however, few have been characterized. In a comparative genetics approach to model such loci in a rat breast cancer model, we previously identified the mammary carcinoma susceptibility locus Mcs1a. We now localize Mcs1a to a critical interval (277 Kb) within a gene desert. Mcs1a reduces mammary carcinoma multiplicity by 50% and acts in a mammary cell-autonomous manner. We developed a megadeletion mouse model, which lacks 535 Kb of sequence containing the Mcs1a ortholog. Global gene expression analysis by RNA-seq revealed that in the mouse mammary gland, the orphan nuclear receptor gene Nr2f1/Coup-tf1 is regulated by Mcs1a. In resistant Mcs1a congenic rats, as compared with susceptible congenic control rats, we found Nr2f1 transcript levels to be elevated in mammary gland, epithelial cells, and carcinoma samples. Chromatin looping over ∼820 Kb of sequence from the Nr2f1 promoter to a strongly conserved element within the Mcs1a critical interval was identified. This element contains a 14 bp indel polymorphism that affects a human-rat-mouse conserved COUP-TF binding motif and is a functional Mcs1a candidate. In both the rat and mouse models, higher Nr2f1 transcript levels are associated with higher abundance of luminal mammary epithelial cells. In both the mouse mammary gland and a human breast cancer global gene expression data set, we found Nr2f1 transcript levels to be strongly anti-correlated to a gene cluster enriched in cell cycle-related genes. We queried 12 large publicly available human breast cancer gene expression studies and found that the median NR2F1 transcript level is consistently lower in 'triple-negative' (ER-PR-HER2-) breast cancers as compared with 'receptor-positive' breast cancers. Our data suggest that the non-protein coding locus Mcs1a regulates Nr2f1, which is a candidate modifier of differentiation, proliferation, and mammary cancer risk.

EBSeq: an empirical Bayes hierarchical model for inference in RNA-seq experiments.

MOTIVATION: Messenger RNA expression is important in normal development and differentiation, as well as in manifestation of disease. RNA-seq experiments allow for the identification of differentially expressed (DE) genes and their corresponding isoforms on a genome-wide scale. However, statistical methods are required to ensure that accurate identifications are made. A number of methods exist for identifying DE genes, but far fewer are available for identifying DE isoforms. When isoform DE is of interest, investigators often apply gene-level (count-based) methods directly to estimates of isoform counts. Doing so is not recommended. In short, estimating isoform expression is relatively straightforward for some groups of isoforms, but more challenging for others. This results in estimation uncertainty that varies across isoform groups. Count-based methods were not designed to accommodate this varying uncertainty, and consequently, application of them for isoform inference results in reduced power for some classes of isoforms and increased false discoveries for others. RESULTS: Taking advantage of the merits of empirical Bayesian methods, we have developed EBSeq for identifying DE isoforms in an RNA-seq experiment comparing two or more biological conditions. Results demonstrate substantially improved power and performance of EBSeq for identifying DE isoforms. EBSeq also proves to be a robust approach for identifying DE genes. AVAILABILITY AND IMPLEMENTATION: An R package containing examples and sample datasets is available at http://www.biostat.wisc.edu/kendzior/EBSEQ/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Comparative RNA-seq analysis in the unsequenced axolotl: the oncogene burst highlights early gene expression in the blastema.

The salamander has the remarkable ability to regenerate its limb after amputation. Cells at the site of amputation form a blastema and then proliferate and differentiate to regrow the limb. To better understand this process, we performed deep RNA sequencing of the blastema over a time course in the axolotl, a species whose genome has not been sequenced. Using a novel comparative approach to analyzing RNA-seq data, we characterized the transcriptional dynamics of the regenerating axolotl limb with respect to the human gene set. This approach involved de novo assembly of axolotl transcripts, RNA-seq transcript quantification without a reference genome, and transformation of abundances from axolotl contigs to human genes. We found a prominent burst in oncogene expression during the first day and blastemal/limb bud genes peaking at 7 to 14 days. In addition, we found that limb patterning genes, SALL genes, and genes involved in angiogenesis, wound healing, defense/immunity, and bone development are enriched during blastema formation and development. Finally, we identified a category of genes with no prior literature support for limb regeneration that are candidates for further evaluation based on their expression pattern during the regenerative process.

Design and measurement of the TE25,10 mode generator.

The high-mode generator is a passive device operating at low power, which is helpful for the mode converter test. It has generally served as the input of the mode converter to evaluate the performance. Here, we realized the design of the TE25,10 mode generator. The multi-section coaxial resonator was designed to improve the TE25,10 mode purity. Two mirrors were used to excite the TE25,10 mode resonance based on the geometric optics. The construction of the TE25,10 mode generator was realized. The purity of the measured TE25,10 mode was 91%, which was in good agreement with the theory.

CD8+ T cell-intrinsic IL-6 signaling promotes resistance to anti-PD-L1 immunotherapy.

Although immune checkpoint inhibitors (ICIs) are established as effective cancer therapies, overcoming therapeutic resistance remains a critical challenge. Here we identify interleukin 6 (IL-6) as a correlate of poor response to atezolizumab (anti-PD-L1) in large clinical trials of advanced kidney, breast, and bladder cancers. In pre-clinical models, combined blockade of PD-L1 and the IL-6 receptor (IL6R) causes synergistic regression of large established tumors and substantially improves anti-tumor CD8+ cytotoxic T lymphocyte (CTL) responses compared with anti-PD-L1 alone. Circulating CTLs from cancer patients with high plasma IL-6 display a repressed functional profile based on single-cell RNA sequencing, and IL-6-STAT3 signaling inhibits classical cytotoxic differentiation of CTLs in vitro. In tumor-bearing mice, CTL-specific IL6R deficiency is sufficient to improve anti-PD-L1 activity. Thus, based on both clinical and experimental evidence, agents targeting IL-6 signaling are plausible partners for combination with ICIs in cancer patients.