Objective To use cell‐based gene signatures to identify patients with systemic lupus erythematous (SLE) in the phase II/III APRIL–SLE and phase IIb ADDRESS II trials most likely to respond to atacicept. Methods A published immune cell deconvolution algorithm based on Affymetrix gene array data was applied to whole blood gene expression from patients entering APRIL‐SLE. Five distinct patient clusters were identified. Patient characteristics, biomarkers, and clinical response to atacicept were assessed per cluster. A modified immune cell deconvolution algorithm was developed based on RNA sequencing data and applied to ADDRESS II data to identify similar patient clusters and their responses. Results Patients in APRIL‐SLE (N = 105) were segregated into the following five clusters (P1‐5) characterized by dominant cell subset signatures: high neutrophils, T helper cells and natural killer (NK) cells (P1), high plasma cells and activated NK cells (P2), high B cells and neutrophils (P3), high B cells and low neutrophils (P4), or high activated dendritic cells, activated NK cells, and neutrophils (P5). Placebo‐ and atacicept‐treated patients in clusters P2,4,5 had markedly higher British Isles Lupus Assessment Group (BILAG) A/B flare rates than those in clusters P1,3, with a greater treatment effect of atacicept on lowering flares in clusters P2,4,5. In ADDRESS II, placebo‐treated patients from P2,4,5 were less likely to be SLE Responder Index (SRI)‐4, SRI‐6, and BILAG‐Based Combined Lupus Assessment responders than those in P1,3; the response proportions again suggested lower placebo effect and a greater treatment differential for atacicept in P2,4,5. Conclusion This exploratory analysis indicates larger differences between placebo‐ and atacicept‐treated patients with SLE in a molecularly defined patient subset.

Systemic lupus erythematosus (SLE) is an autoimmune disease affecting multiple organ systems. Many investigational agents have failed or shown only modest effects when added to standard of care (SoC) therapy in placebo-controlled trials, and only two therapies have been approved for SLE in the last 60 years. Clinical trial outcomes have shown discordance in drug effects between clinical endpoints. Herein, we characterized longitudinal disease activity in the SLE population and the sources of variability by developing a latent disease trajectory model for SLE component endpoints (Systemic Lupus Erythematosus Disease Activity Index [SLEDAI], Physician's Global Assessment [PGA], British Isles Lupus Assessment Group Index [BILAG]) and composite endpoints (Systemic Lupus Erythematosus Responder Index [SRI], BILAG-based Composite Lupus Assessment [BICLA], and Lupus Low Disease Activity State [LLDAS]) using patient-level historical SoC data from nine phase II and III studies. Across all endpoints, in predictions up to 52 weeks from the final disease trajectory model, the following baseline covariates were associated with a greater decrease in SLE disease activity and higher response to placebo + SoC: Hispanic ethnicity from Central/South America, absence of hypocomplementemia, recent SLE diagnosis, and high baseline disease activity score using SLEDAI and BILAG separately. No discernible differences were observed in the trajectory of response to placebo + SoC across different SoC medications (antimalarial and immunosuppressant such as mycophenolate, methotrexate, and azathioprine). Across all endpoints, disease trajectory showed no difference in Asian versus non-Asian patients, supporting Asia-inclusive global SLE drug development. These results describe the first population approach to support a model-informed drug development framework in SLE.

The gene regulatory network (GRN) of human cells encodes mechanisms to ensure proper functioning. However, if this GRN is dysregulated, the cell may enter into a disease state such as cancer. Understanding the GRN as a system can therefore help identify novel mechanisms underlying disease, which can lead to new therapies. To deduce regulatory interactions relevant to cancer, we applied a recent computational inference framework to data from perturbation experiments in squamous carcinoma cell line A431. GRNs were inferred using several methods, and the false discovery rate was controlled by the NestBoot framework. We developed a novel approach to assess the predictiveness of inferred GRNs against validation data, despite the lack of a gold standard. The best GRN was significantly more predictive than the null model, both in cross-validated benchmarks and for an independent dataset of the same genes under a different perturbation design. The inferred GRN captures many known regulatory interactions central to cancer-relevant processes in addition to predicting many novel interactions, some of which were experimentally validated, thus providing mechanistic insights that are useful for future cancer research.

Soybean aphids (Aphis glycines Matsumura) are specialized insects that feed on soybean (Glycine max) phloem sap. Transcriptome analyses have shown that resistant soybean plants mount a fast response that limits aphid feeding and population growth. Conversely, defense responses in susceptible plants are slower and it is hypothesized that aphids block effective defenses in the compatible interaction. Unlike other pests, aphids can colonize plants for long periods of time; yet the effect on the plant transcriptome after long-term aphid feeding has not been analyzed for any plant-aphid interaction. We analyzed the susceptible and resistant (Rag1) transcriptome response to aphid feeding in soybean plants colonized by aphids (biotype 1) for 21 days. We found a reduced resistant response and a low level of aphid growth on Rag1 plants, while susceptible plants showed a strong response consistent with pattern-triggered immunity. GO-term analyses identified chitin regulation as one of the most overrepresented classes of genes, suggesting that chitin could be one of the hemipteran-associated molecular pattern that triggers this defense response. Transcriptome analyses also indicated the phenylpropanoid pathway, specifically isoflavonoid biosynthesis, was induced in susceptible plants in response to long-term aphid feeding. Metabolite analyses corroborated this finding. Aphid-treated susceptible plants accumulated daidzein, formononetin, and genistein, although glyceollins were present at low levels in these plants. Choice experiments indicated that daidzein may have a deterrent effect on aphid feeding. Mass spectrometry imaging showed these isoflavones accumulate likely in the mesophyll cells or epidermis and are absent from the vasculature, suggesting that isoflavones are part of a non-phloem defense response that can reduce aphid feeding. While it is likely that aphid can initially block defense responses in compatible interactions, it appears that susceptible soybean plants can eventually mount an effective defense in response to long-term soybean aphid colonization.

Background The Phase II/III APRIL-SLE study evaluated the safety and efficacy of atacicept in systemic lupus erythematosus (SLE). The goal of this post-hoc analysis was to use cell-based gene signatures on the APRIL-SLE gene expression data to identify clusters of patients with potential to flare and to assess for difference in treatment effect of atacicept vs placebo. Methods A published immune cell deconvolution algorithm was applied to whole-blood gene expression data from APRIL-SLE to identify relative proportions of 17 immune cell types. Patients were then grouped into clusters based on these immune cell profiles using a k-medoid clustering algorithm, and were compared to each other based on patient characteristics, biomarkers and clinical efficacy. In addition, the baseline expression and change in expression of putative APRIL-responder genes were compared among the clusters. APRIL-responder genes were identified by combining differential expression results from the APRIL-SLE study (Week 52 vs. Day 1 randomization) and tabalumab Phase III studies (Week 52 vs. Baseline; GSE88887). Results Patient gene expression data (N=105; Placebo: N=30; atacicept 75 mg: N=40; atacicept 150 mg N=35) was used to group patients into 5 main clusters (P1-P5) by predominant characteristic cells: P1, T helper cells; P2, plasma cells; P3, neutrophils and B cells; P4, B cells; P5, activated dendritic cells. Patients in P2 and P5 were more likely to have positive anti-dsDNA antibodies (≥30 IU/ml) and elevated BLyS (≥1.6 ng/ml), as well as high IFN gene signature in the blood. Patients in P2 were more likely to have low complement C3 and C4 levels. In P2, P4, and P5 clusters the flare rate in the placebo group was significantly higher than in P1 and P3. In P2 and P4, atacicept 150mg treatment group showed delayed time to flare and reduced flare rate as compared with the placebo group. A comparison of differentially-expressed genes from clinical studies of SLE patients on atacicept (targets BLyS & APRIL) vs tabalumab (targets BLyS) revealed possible APRIL-responder genes: SDC1, PARM1 and MZB1. These genes have a higher baseline expression in the P2 and P4 compared to other clusters. SDC1 was reduced more in P2, P4, and P5 after atacicept treatment, while PARM1 and MZB1 decreased after atacicept treatment in P2 and P4. Conclusions These post-hoc analyses revealed different subsets of SLE patients based on their molecular profiles, which identified patient subsets that might have differential treatment effect of atacicept vs placebo, and provided insights into potential mechanisms of flare. Funding Source(s): Merck KGaA, Darmstadt, Germany

Background Pharmacodynamic biomarkers are becoming increasingly valuable for assessing drug activity and target modulation in clinical trials. However, identifying quality biomarkers is challenging due to the increasing volume and heterogeneity of relevant data describing the biological networks that underlie disease mechanisms. A biological pathway network typically includes entities (e.g. genes, proteins and chemicals/drugs) as well as the relationships between these and is typically curated or mined from structured databases and textual co-occurrence data. We propose a hybrid Natural Language Processing and directed relationships-based network analysis approach using IBM Watson for Drug Discovery to rank all human genes and identify potential candidate biomarkers, requiring only an initial determination of a specific target-disease relationship. Methods Through natural language processing of scientific literature, Watson for Drug Discovery creates a network of semantic relationships between biological concepts such as genes, drugs, and diseases. Using Bruton’s tyrosine kinase as a case study, Watson for Drug Discovery’s automatically extracted relationship network was compared with a prominent manually curated physical interaction network. Additionally, potential biomarkers for Bruton’s tyrosine kinase inhibition were predicted using a matrix factorization approach and subsequently compared with expert-generated biomarkers. Results Watson’s natural language processing generated a relationship network matching 55 (86%) genes upstream of BTK and 98 (95%) genes downstream of Bruton’s tyrosine kinase in a prominent manually curated physical interaction network. Matrix factorization analysis predicted 11 of 13 genes identified by Merck subject matter experts in the top 20% of Watson for Drug Discovery’s 13,595 ranked genes, with 7 in the top 5%. Conclusion Taken together, these results suggest that Watson for Drug Discovery’s automatic relationship network identifies the majority of upstream and downstream genes in biological pathway networks and can be used to help with the identification and prioritization of pharmacodynamic biomarker evaluation, accelerating the early phases of disease hypothesis generation.

Background: The Phase 2/3 APRIL-SLE study evaluated the safety and efficacy of atacicept, a dual inhibitor of the B lymphocyte stimulator (BLyS) and a proliferation-inducing ligand (APRIL), in systemic lupus erythematosus (SLE). Objectives: The goal of this post-hoc analysis was to use cell-based gene signatures on gene expression data from the APRIL-SLE study to identify clusters of patients with potential to flare and to assess clusters for differences in treatment effects of atacicept vs placebo. Methods: A published immune cell deconvolution algorithm (Abbas et al, 2009) was applied to whole-blood gene expression data from APRIL-SLE patients to identify relative proportions of 17 immune cell types. Patients were then grouped into clusters based on these immune cell profiles using a k-medoid clustering algorithm and were compared to each other based on patient characteristics, biomarkers and clinical efficacy. In addition, baseline expression and change in expression of putative APRIL-responder genes were compared among clusters. APRIL-responder genes were identified by combining differential expression results from the APRIL-SLE study (Week 52 vs Day 1 randomization) and tabalumab (targets BLyS only) Phase 3 studies (Week 52 vs baseline; GSE88887). Results: Patient gene expression data (N=105; placebo, n=30; atacicept 75 mg, n=40; atacicept 150 mg, n=35) were used to group patients into five main clusters (P1-P5) by predominant characteristic cells: P1, T helper cells; P2, plasma cells; P3, neutrophils and B cells; P4, B cells; P5, activated dendritic cells. Patients in P2 and P5 were more likely to have positive anti-dsDNA antibodies (≥30 IU/ml), elevated BLyS; ≥1.6 ng/ml, and high interferon gene signature in the blood than other those in other clusters. Patients in P2 were most likely to have low complement C3 and C4 levels. Placebo-group flare rates in P2 (100%), P4 (100%) and P5 (83%) were markedly higher than in P1 (33%) and P3 (29%). In P2, P4, and P5 the median time-to-flare was much lower with placebo (85, 98.5, and 115.5 days, respectively) than with atacicept 150 mg (over 364 days for all three clusters). A comparison of differentially-expressed genes from clinical studies of SLE patients treated with atacicept and tabalumab revealed possible APRIL-responder genes: SDC1, PARM1 and MZB1. These genes had a higher baseline expression in P2 and P4 compared with other clusters. SDC1 was reduced from baseline more in P2, P4, and P5 after atacicept treatment, while PARM1 and MZB1 decreased after atacicept treatment in P2 and P4. Conclusion: These post-hoc analyses revealed different subsets of SLE patients based on their molecular profiles. Atacicept may have different treatment effects in the identified patient subsets vs placebo, providing insights into potential mechanisms of flare in SLE. References: [1] Abbas AR, Wolslegel K, Seshasayee D, Modrusan Z, Clark HF. Deconvolution of blood microarray data identifies cellular activation patterns in systemic lupus erythematosus. PLoS ONE. 2009;4(7):e6098. Disclosure of Interests: Eileen Samy Employee of: Current employees of EMD Serono, Matthew Studham Employee of: Current employees of EMD Serono, Amy Kao Employee of: Current employees of EMD Serono, Philipp Haselmayer Employee of: Current employee of Merck KGaA, Peter Chang Employee of: Current employees of EMD Serono, P. Alexander Rolfe Employee of: Current employees of EMD Serono, David Wofsy Consultant for: GlaxoSmithKline – Member, data safety monitoring board Novartis – Member, data safety monitoring board Celgene – member, scientific advisory board, Julie DeMartino Shareholder of: Shares in Merck & Co, Employee of: Current employees of EMD Serono; Past employee of Merck & Co. & Roche, Robert Townsend Employee of: Current employees of EMD Serono

The NUDIX enzymes are involved in cellular metabolism and homeostasis, as well as mRNA processing. Although highly conserved throughout all organisms, their biological roles and biochemical redundancies remain largely unclear. To address this, we globally resolve their individual properties and inter-relationships. We purify 18 of the human NUDIX proteins and screen 52 substrates, providing a substrate redundancy map. Using crystal structures, we generate sequence alignment analyses revealing four major structural classes. To a certain extent, their substrate preference redundancies correlate with structural classes, thus linking structure and activity relationships. To elucidate interdependence among the NUDIX hydrolases, we pairwise deplete them generating an epistatic interaction map, evaluate cell cycle perturbations upon knockdown in normal and cancer cells, and analyse their protein and mRNA expression in normal and cancer tissues. Using a novel FUSION algorithm, we integrate all data creating a comprehensive NUDIX enzyme profile map, which will prove fundamental to understanding their biological functionality.

A key question in network inference, that has not been properly answered, is what accuracy can be expected for a given biological dataset and inference method. We present GeneSPIDER - a Matlab package for tuning, running, and evaluating inference algorithms that allows independent control of network and data properties to enable data-driven benchmarking. GeneSPIDER is uniquely suited to address this question by first extracting salient properties from the experimental data and then generating simulated networks and data that closely match these properties. It enables data-driven algorithm selection, estimation of inference accuracy from biological data, and a more multifaceted benchmarking. Included are generic pipelines for the design of perturbation experiments, bootstrapping, analysis of linear dependence, sample selection, scaling of SNR, and performance evaluation. With GeneSPIDER we aim to move the goal of network inference benchmarks from simple performance measurement to a deeper understanding of how the accuracy of an algorithm is determined by different combinations of network and data properties.

The soybean aphid (Aphis glycines) is one of the main insect pests of soybean (Glycine max) worldwide. Genomics approaches have provided important data on transcriptome changes, both in the insect and in the plant, in response to the plant-aphid interaction. However, the difficulties to transform soybean and to rear soybean aphid on artificial media have hindered our ability to systematically test the function of genes identified by those analyses as mediators of plant resistance to the insect. An efficient approach to produce transgenic soybean material is the production of transformed hairy roots using Agrobacterium rhizogenes; however, soybean aphids colonize leaves or stems and thus this approach has not been utilized. Here, we developed a hairy root system that allowed effective aphid feeding. We show that this system supports aphid performance similar to that observed in leaves. The use of hairy roots to study plant resistance is validated by experiments showing that roots generated from cotyledons of resistant lines carrying the Rag1 or Rag2 resistance genes are also resistant to aphid feeding, while related susceptible lines are not. Our results demonstrate that hairy roots are a good system to study soybean aphid-soybean interactions, providing a quick and effective method that could be used for functional analysis of the resistance response to this insect.

A range of tools are available to model, simulate and analyze gene regulatory networks (GRNs). However, these tools provide limited ability to control network topology, system dynamics, design of e ...

Statistical regularisation methods such as LASSO and related L1 regularised regression methods are commonly used to construct models of gene regulatory networks. Although they can theoretically infer the correct network structure, they have been shown in practice to make errors, i.e. leave out existing links and include non-existing links. We show that L1 regularisation methods typically produce a poor network model when the analysed data are ill-conditioned, i.e. the gene expression data matrix has a high condition number, even if it contains enough information for correct network inference. However, the correct structure of network models can be obtained for informative data, data with such a signal to noise ratio that existing links can be proven to exist, when these methods fail, by using least-squares regression and setting small parameters to zero, or by using robust network inference, a recent method taking the intersection of all non-rejectable models. Since available experimental data sets are generally ill-conditioned, we recommend to check the condition number of the data matrix to avoid this pitfall of L1 regularised inference, and to also consider alternative methods.

MOTIVATION:Gene regulatory network (GRN) inference reveals the influences genes have on one another in cellular regulatory systems. If the experimental data are inadequate for reliable inference of the network, informative priors have been shown to improve the accuracy of inferences.RESULTS:This study explores the potential of undirected, confidence-weighted networks, such as those in functional association databases, as a prior source for GRN inference. Such networks often erroneously indicate symmetric interaction between genes and may contain mostly correlation-based interaction information. Despite these drawbacks, our testing on synthetic datasets indicates that even noisy priors reflect some causal information that can improve GRN inference accuracy. Our analysis on yeast data indicates that using the functional association databases FunCoup and STRING as priors can give a small improvement in GRN inference accuracy with biological data.

Gene regulatory network inference (that is, determination of the regulatory interactions between a set of genes) provides mechanistic insights of central importance to research in systems biology. Most contemporary network inference methods rely on a sparsity/regularization coefficient, which we call ζ (zeta), to determine the degree of sparsity of the network estimates, that is, the total number of links between the nodes. However, they offer little or no advice on how to select this sparsity coefficient, in particular, for biological data with few samples. We show that an empty network is more accurate than estimates obtained for a poor choice of ζ. In order to avoid such poor choices, we propose a method for optimization of ζ, which maximizes the accuracy of the inferred network for any sparsity-dependent inference method and data set. Our procedure is based on leave-one-out cross-optimization and selection of the ζ value that minimizes the prediction error. We also illustrate the adverse effects of noise, few samples, and uninformative experiments on network inference as well as our method for optimization of ζ. We demonstrate that our ζ optimization method for two widely used inference algorithms--Glmnet and NIR--gives accurate and informative estimates of the network structure, given that the data is informative enough.

BACKGROUND:Phytohormones mediate plant defense responses to pests and pathogens. In particular, the hormones jasmonic acid, ethylene, salicylic acid, and abscisic acid have been shown to dictate and fine-tune defense responses, and identification of the phytohormone components of a particular defense response is commonly used to characterize it. Identification of phytohormone regulation is particularly important in transcriptome analyses. Currently there is no computational tool to determine the relative activity of these hormones that can be applied to transcriptome analyses in soybean.FINDINGS:We developed a pathway analysis method that provides a broad measure of the activation or suppression of individual phytohormone pathways based on changes in transcript expression of pathway-related genes. The magnitude and significance of these changes are used to determine a pathway score for a phytohormone for a given comparison in a microarray experiment. Scores for individual hormones can then be compared to determine the dominant phytohormone in a given defense response. To validate this method, it was applied to publicly available data from previous microarray experiments that studied the response of soybean plants to Asian soybean rust and soybean cyst nematode. The results of the analyses for these experiments agreed with our current understanding of the role of phytohormones in these defense responses.CONCLUSIONS:This method is useful in providing a broad measure of the relative induction and suppression of soybean phytohormones during a defense response. This method could be used as part of microarray studies that include individual transcript analysis, gene set analysis, and other methods for a comprehensive defense response characterization.

The soybean aphid (Aphis glycines) is a major phloem-feeding pest of soybean (Glycine max). A. glycines feeding can cause the diversion of photosynthates and transmission of plant viruses, resulting in significant yield losses. In this study, we used oligonucleotide microarrays to characterize the long-term transcriptional response to soybean aphid colonization of two related soybean cultivars, one with the Rag1 aphid-resistance gene and one aphid-susceptible cultivar (without Rag1). Transcriptome profiles were determined after 1 and 7 days of aphid infestation. Our results revealed a susceptible response involving hundreds of transcripts, whereas only one transcript changed in the resistant response to aphids. This nonexistent resistance response might be explained by the fact that many defense-related transcripts are constitutively expressed in resistant plants, whereas these same genes are activated in susceptible plants only during aphid infestation. Analysis of phytohormone-related transcripts in the susceptible response showed different hormone profiles for the two time points, and suggest that aphids are able to suppress hormone signals in susceptible plants. A significant activation of abscissic acid, normally associated with abiotic stress responses, at day 7, might be a decoy strategy implemented by the aphid to suppress effective salicylic acid- and jasmonate-related defenses.

The only well-characterized gene controlling soybean aphid (SBA) resistance is Mi-1.2, a tomato gene that also confers resistance to root-knot nematodes (RKN). Based on similarities between Mi-1.2 and Rag1, which produces a strong antibiosis-type resistance, the authors hypothesized that Rag1 could also provide resistance to nematodes. They evaluated two soybean lines, one carrying the Rag1 gene, and one with no resistance gene, for susceptibility or resistance to soybean cyst nematode (SCN) and RKN. Accepted for publication 4 March 2009. Published 1 April 2009.