Search

PDF file - 365K, The MEK inhibitor-resistant cells re-express pERK at later time point after treatment.

PDF file - 2608K, RAS mRNA is not affected by DDX43 depletion.

PDF file - 766K, DDX43 is silenced with a second siRNA.

PDF file - 39K, The selumetinib-resistant UM cells are also resistant to other MEK inhibitors.

PDF file - 76K, Examples of negative results with the NanoString assay for kinase fusions.

Purpose: Gastric cancer may be subdivided into 3 distinct subtypes—proximal, diffuse, and distal gastric cancer—based on histopathologic and anatomic criteria. Each subtype is associated with unique epidemiology. Our aim is to test the hypothesis that these distinct gastric cancer subtypes may also be distinguished by gene expression analysis.Experimental Design: Patients with localized gastric adenocarcinoma being screened for a phase II preoperative clinical trial (National Cancer Institute, NCI #5917) underwent endoscopic biopsy for fresh tumor procurement. Four to 6 targeted biopsies of the primary tumor were obtained. Macrodissection was carried out to ensure more than 80% carcinoma in the sample. HG-U133A GeneChip (Affymetrix) was used for cDNA expression analysis, and all arrays were processed and analyzed using the Bioconductor R-package.Results: Between November 2003 and January 2006, 57 patients were screened to identify 36 patients with localized gastric cancer who had adequate RNA for expression analysis. Using supervised analysis, we built a classifier to distinguish the 3 gastric cancer subtypes, successfully classifying each into tightly grouped clusters. Leave-one-out cross-validation error was 0.14, suggesting that more than 85% of samples were classified correctly. Gene set analysis with the false discovery rate set at 0.25 identified several pathways that were differentially regulated when comparing each gastric cancer subtype to adjacent normal stomach.Conclusions: Subtypes of gastric cancer that have epidemiologic and histologic distinctions are also distinguished by gene expression data. These preliminary data suggest a new classification of gastric cancer with implications for improving our understanding of disease biology and identification of unique molecular drivers for each gastric cancer subtype. Clin Cancer Res; 17(9); 2693–701. ©2011 AACR.

PDF file - 308K, Complete list of target sequences for NanoString probes

Background: The mutually exclusive pattern of the major driver oncogenes in lung cancer suggests that other mutually exclusive oncogenes exist. We conducted a systematic search for tyrosine kinase fusions by screening all tyrosine kinases for aberrantly high RNA expression levels of the 3′ kinase domain (KD) exons relative to more 5′ exons.Methods: We studied 69 patients (including five never smokers and 64 current or former smokers) with lung adenocarcinoma negative for all major mutations in KRAS, EGFR, BRAF, MEK1, HER2, and for ALK fusions (termed “pan-negative”). A NanoString-based assay was designed to query the transcripts of 90 tyrosine kinases at two points: 5′ to the KD and within the KD or 3′ to it. Tumor RNAs were hybridized to the NanoString probes and analyzed for outlier 3′ to 5′ expression ratios. Presumed novel fusion events were studied by rapid amplification of cDNA ends (RACE) and confirmatory reverse transcriptase PCR (RT-PCR) and FISH.Results: We identified one case each of aberrant 3′ to 5′ ratios in ROS1 and RET. RACE isolated a GOPC-ROS1 (FIG-ROS1) fusion in the former and a KIF5B-RET fusion in the latter, both confirmed by RT-PCR. The RET rearrangement was also confirmed by FISH. The KIF5B-RET patient was one of only five never smokers in this cohort.Conclusion: The KIF5B-RET fusion defines an additional subset of lung cancer with a potentially targetable driver oncogene enriched in never smokers with “pan-negative” lung adenocarcinomas. We also report in lung cancer the GOPC-ROS1 fusion originally discovered and characterized in a glioma cell line. Clin Cancer Res; 18(24); 6599–608. ©2012 AACR.

PDF file - 32K, Clinicopathological characteristics of study cohort

Supplementary Methods - PDF file 158K, Supplemental methods contain both techniques and further data strengthening the manuscript. These include our Analysis of Breast Cancer Microarray Datasets, Generation of LDH-A knockdown and control cell lines and In vitro assays, and in vivo methods (MR measurements)

Purpose: To investigate the relationship between lactate dehydrogenase A (LDH-A) expression, lactate concentration, cell metabolism, and metastases in murine 4T1 breast tumors.Experimental Design: Inhibition of LDH-A expression and protein levels were achieved in a metastatic breast cancer cell line (4T1) using short hairpin RNA (shRNA) technology. The relationship between tumor LDH-A protein levels and lactate concentration (measured by magnetic resonance spectroscopic imaging, MRSI) and metastases was assessed.Results: LDH-A knockdown cells (KD9) showed a significant reduction in LDH-A protein and LDH activity, less acid production, decreased transwell migration and invasion, lower proliferation, reduced glucose consumption and glycolysis, and increase in oxygen consumption, reactive oxygen species (ROS), and cellular ATP levels, compared with control (NC) cells cultured in 25 mmol/L glucose. In vivo studies showed lower lactate levels in KD9, KD5, and KD317 tumors than in NC or 4T1 wild-type tumors (P < 0.01), and a linear relationship between tumor LDH-A protein expression and lactate concentration. Metastases were delayed and primary tumor growth rate decreased.Conclusions: We show for the first time that LDH-A knockdown inhibited the formation of metastases, and was accompanied by in vivo changes in tumor cell metabolism. Lactate MRSI can be used as a surrogate to monitor targeted inhibition of LDH-A in a preclinical setting and provides a noninvasive imaging strategy to monitor LDH-A–targeted therapy. This imaging strategy can be translated to the clinic to identify and monitor patients who are at high risk of developing metastatic disease. Clin Cancer Res; 19(18); 5158–69. ©2013 AACR.

Supplementary Figures - PDF file 259K, Figure S1. Comparison of 4T1-WT, NC and KD9 cell lines. (A) Oxylite Oxygen Consumption Rate; (B) Reactive oxygen species (ROS); (C) mitochondria level; (D) Primary tumor (~ 100 mm3) lactate concentration. Figure S2. Histology of small (100 mm3) primary NC and KD9 tumors (A). Representative small tumors show similar features: a few small areas of necrosis (H&E) and reduced blood flow (Hoechst). Scale bar represents 5 mm. Cell Mitochondria (B). NC and KD9 cells were grown on the cover slips and stained for mitochondria by Mitotracker red CMXRos. Fig. S3. Primary orthotopic tumor growth and development of lung metastases. Three weeks after tumor inoculation, the mice were sacrificed and the primary tumor volume and weight, as well as lung weight (reflecting the presence of lung metastases) were measured in NC (n=6) and KD9 (n=6) tumor bearing animals Fig. S4. Western blot analysis of LDH-A expression: The absence of puromycin in the culture media results in a reappearance of LDH-A expression in KD9 cells growing in vitro (columns 3 and 5)

Supplementary Figure S1; Supplementary Table S1.

Liposarcoma remains the most common mesenchymal cancer, with a mortality rate of 60% among patients with this disease. To address the present lack of therapeutic options, we embarked upon a study of microRNA (miRNA) expression alterations associated with liposarcomagenesis with the goal of exploiting differentially expressed miRNAs and the gene products they regulate as potential therapeutic targets. MicroRNA expression was profiled in samples of normal adipose tissue, well-differentiated liposarcoma, and dedifferentiated liposarcoma by both deep sequencing of small RNA libraries and hybridization-based Agilent microarrays. The expression profiles discriminated liposarcoma from normal adipose tissue and well differentiated from dedifferentiated disease. We defined over 40 miRNAs that were dysregulated in dedifferentiated liposarcomas in both the sequencing and the microarray analysis. The upregulated miRNAs included two cancer-associated species (miR-21 and miR-26a), and the downregulated miRNAs included two species that were highly abundant in adipose tissue (miR-143 and miR-145). Restoring miR-143 expression in dedifferentiated liposarcoma cells inhibited proliferation, induced apoptosis, and decreased expression of BCL2, topoisomerase 2A, protein regulator of cytokinesis 1 (PRC1), and polo-like kinase 1 (PLK1). The downregulation of PRC1 and its docking partner PLK1 suggests that miR-143 inhibits cytokinesis in these cells. In support of this idea, treatment with a PLK1 inhibitor potently induced G2–M growth arrest and apoptosis in liposarcoma cells. Taken together, our findings suggest that miR-143 re-expression vectors or selective agents directed at miR-143 or its targets may have therapeutic value in dedifferentiated liposarcoma. Cancer Res; 71(17); 5659–69. ©2011 AACR.

Preeclampsia and gestational hypertensive disorders (GHD) are common complications of pregnancy that adversely affect maternal and offspring health, often with long-term consequences. High BMI, advanced age, and pre-existing conditions are known risk factors for GHD. Yet, assessing a woman’s risk of GHD based on only these characteristics needs to be reevaluated in order to identify at-risk women, facilitate early diagnosis, and implement lifestyle recommendations. This study demonstrates that a risk score developed with machine learning from the case-control genetics dataset can be used as an early screening test for GHD. We further confirm BMI as a risk factor for GHD and investigate a relationship between GHD and genetically constructed anthropometric measures and biomarkers. Our results show that polygenic risk score can be used as an early screening tool that, together with other known risk factors and medical history, would assist in identifying women at higher risk of GHD before its onset to enable stratification of patients into low-risk and high-risk groups for monitoring and preventative programs to mitigate the risks.

Purpose The major causes of mortality after allogeneic hematopoietic-cell transplantation (allo-HCT) are relapse, graft-versus-host disease (GVHD), and infection. We have reported previously that alterations in the intestinal flora are associated with GVHD, bacteremia, and reduced overall survival after allo-HCT. Because intestinal bacteria are potent modulators of systemic immune responses, including antitumor effects, we hypothesized that components of the intestinal flora could be associated with relapse after allo-HCT. Methods The intestinal microbiota of 541 patients admitted for allo-HCT was profiled by means of 16S ribosomal sequencing of prospectively collected stool samples. We examined the relationship between abundance of microbiota species or groups of related species and relapse/progression of disease during 2 years of follow-up time after allo-HCT by using cause-specific proportional hazards in a retrospective discovery-validation cohort study. Results Higher abundance of a bacterial group composed mostly of Eubacterium limosum in the validation set was associated with a decreased risk of relapse/progression of disease (hazard ratio [HR], 0.82 per 10-fold increase in abundance; 95% CI, 0.71 to 0.95; P = .009). When the patients were categorized according to presence or absence of this bacterial group, presence also was associated with less relapse/progression of disease (HR, 0.52; 95% CI, 0.31 to 0.87; P = .01). The 2-year cumulative incidences of relapse/progression among patients with and without this group of bacteria were 19.8% and 33.8%, respectively. These associations remained significant in multivariable models and were strongest among recipients of T-cell–replete allografts. Conclusion We found associations between the abundance of a group of bacteria in the intestinal flora and relapse/progression of disease after allo-HCT. These might serve as potential biomarkers or therapeutic targets to prevent relapse and improve survival after allo-HCT.

Genetic underpinnings of facial aging are still largely unknown. In this study, we leverage the statistical power of large-scale data from the UK Biobank and perform insilico analysis of genome-wide self-perceived facial aging. Functional analysis reveals significant over-representation of skin pigmentation and immune related pathways that are correlated with facial aging. For males, hair loss is one of the top categories that is highly significantly over-represented in the genetics data associated with self-reported facial aging. Our analysis confirms that genes coding for the extracellular matrix play important roles in aging. Overall, our results provide evidence that while somewhat biased, large-scale self-reported data on aging can be utilized for extracting useful insights into underlying biology, provide candidate skin aging biomarkers, and advance anti-aging skincare.

The last decade has seen a marked rise in the use of cancer tissues obtained from research autopsies. Such resources have been invaluable for studying cancer evolution or the mechanisms of therapeutic resistance to targeted therapies. Degradation of biomolecules is a potential challenge to usage of cancer tissues obtained in the post-mortem setting and remains incompletely studied. We analysed the nucleic acid quality in 371 different frozen tissue samples collected from 80 patients who underwent a research autopsy, including eight normal tissue types, primary and metastatic tumors. Our results indicate that RNA integrity number (RIN) of normal tissues decline with the elongation of post-mortem interval (PMI) in a tissue-type specific manner. Unlike normal tissues, the RNA quality of cancer tissues is highly variable with respect to post-mortem interval. The kinetics of DNA damage also has tissue type-specific features. Moreover, while DNA degradation is an indicator of low RNA quality, the converse is not true. Finally, we show that despite RIN values as low as 5.0, robust data can be obtained by RNA sequencing that reliably discriminates expression signatures.

The composition of the intestinal microbiota influences the development of inflammatory disorders. However, associating inflammatory diseases with specific microbial members of the microbiota is challenging, because clinically detectable inflammation and its treatment can alter the microbiota’s composition. Immunologic checkpoint blockade with ipilimumab, a monoclonal antibody that blocks cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) signalling, is associated with new-onset, immune-mediated colitis. Here we conduct a prospective study of patients with metastatic melanoma undergoing ipilimumab treatment and correlate the pre-inflammation faecal microbiota and microbiome composition with subsequent colitis development. We demonstrate that increased representation of bacteria belonging to the Bacteroidetes phylum is correlated with resistance to the development of checkpoint-blockade-induced colitis. Furthermore, a paucity of genetic pathways involved in polyamine transport and B vitamin biosynthesis is associated with an increased risk of colitis. Identification of these biomarkers may enable interventions to reduce the risk of inflammatory complications following cancer immunotherapy.

Despite a growing understanding of the link between intestinal inflammation and resident gut microbes, longitudinal studies of human flora before initial onset of intestinal inflammation have not been reported. Here, we demonstrate in murine and human recipients of allogeneic bone marrow transplantation (BMT) that intestinal inflammation secondary to graft-versus-host disease (GVHD) is associated with major shifts in the composition of the intestinal microbiota. The microbiota, in turn, can modulate the severity of intestinal inflammation. In mouse models of GVHD, we observed loss of overall diversity and expansion of Lactobacillales and loss of Clostridiales. Eliminating Lactobacillales from the flora of mice before BMT aggravated GVHD, whereas reintroducing the predominant species of Lactobacillus mediated significant protection against GVHD. We then characterized gut flora of patients during onset of intestinal inflammation caused by GVHD and found patterns mirroring those in mice. We also identified increased microbial chaos early after allogeneic BMT as a potential risk factor for subsequent GVHD. Together, these data demonstrate regulation of flora by intestinal inflammation and suggest that flora manipulation may reduce intestinal inflammation and improve outcomes for allogeneic BMT recipients.

Angiosarcomas (ASs) represent a heterogeneous group of malignant vascular tumors that may occur spontaneously as primary tumors or secondarily after radiation therapy or in the context of chronic lymphedema. Most secondary ASs have been associated with MYC oncogene amplification, whereas the role of MYC abnormalities in primary AS is not well defined. Twenty-two primary and secondary ASs were analyzed by array-comparative genomic hybridization (aCGH) and by deep sequencing of small RNA libraries. By aCGH and subsequently confirmed by fluorescence in situ hybridization, MYC amplification was identified in three out of six primary tumors and in 8 out of 12 secondary AS. We have also found MAML1 as a new potential oncogene in MYC-amplified AS. Significant upregulation of the miR-17-92 cluster was observed in MYC-amplified AS compared to AS lacking MYC amplification and the control group (other vascular tumors, nonvascular sarcomas). Moreover, MYC-amplified ASs were associated with a significantly lower expression of thrombospondin-1 (THBS1) than AS without MYC amplification or controls. Altogether, our study implicates MYC amplification not only in the pathogenesis of secondary AS but also in a subset of primary AS. Thus, MYC amplification may play a crucial role in the angiogenic phenotype of AS through upregulation of the miR-17-92 cluster, which subsequently downregulates THBS1, a potent endogenous inhibitor of angiogenesis. © 2012 Wiley Periodicals, Inc.

Recurrent mutations in isocitrate dehydrogenase 1 (IDH1) and IDH2 have been identified in gliomas, acute myeloid leukaemias (AML) and chondrosarcomas, and share a novel enzymatic property of producing 2-hydroxyglutarate (2HG) from α-ketoglutarate. Here we report that 2HG-producing IDH mutants can prevent the histone demethylation that is required for lineage-specific progenitor cells to differentiate into terminally differentiated cells. In tumour samples from glioma patients, IDH mutations were associated with a distinct gene expression profile enriched for genes expressed in neural progenitor cells, and this was associated with increased histone methylation. To test whether the ability of IDH mutants to promote histone methylation contributes to a block in cell differentiation in non-transformed cells, we tested the effect of neomorphic IDH mutants on adipocyte differentiation in vitro. Introduction of either mutant IDH or cell-permeable 2HG was associated with repression of the inducible expression of lineage-specific differentiation genes and a block to differentiation. This correlated with a significant increase in repressive histone methylation marks without observable changes in promoter DNA methylation. Gliomas were found to have elevated levels of similar histone repressive marks. Stable transfection of a 2HG-producing mutant IDH into immortalized astrocytes resulted in progressive accumulation of histone methylation. Of the marks examined, increased H3K9 methylation reproducibly preceded a rise in DNA methylation as cells were passaged in culture. Furthermore, we found that the 2HG-inhibitable H3K9 demethylase KDM4C was induced during adipocyte differentiation, and that RNA-interference suppression of KDM4C was sufficient to block differentiation. Together these data demonstrate that 2HG can inhibit histone demethylation and that inhibition of histone demethylation can be sufficient to block the differentiation of non-transformed cells.

Purpose: Gastric cancer may be subdivided into 3 distinct subtypes—proximal, diffuse, and distal gastric cancer—based on histopathologic and anatomic criteria. Each subtype is associated with unique epidemiology. Our aim is to test the hypothesis that these distinct gastric cancer subtypes may also be distinguished by gene expression analysis. Experimental Design: Patients with localized gastric adenocarcinoma being screened for a phase II preoperative clinical trial (National Cancer Institute, NCI #5917) underwent endoscopic biopsy for fresh tumor procurement. Four to 6 targeted biopsies of the primary tumor were obtained. Macrodissection was carried out to ensure more than 80% carcinoma in the sample. HG-U133A GeneChip (Affymetrix) was used for cDNA expression analysis, and all arrays were processed and analyzed using the Bioconductor R-package. Results: Between November 2003 and January 2006, 57 patients were screened to identify 36 patients with localized gastric cancer who had adequate RNA for expression analysis. Using supervised analysis, we built a classifier to distinguish the 3 gastric cancer subtypes, successfully classifying each into tightly grouped clusters. Leave-one-out cross-validation error was 0.14, suggesting that more than 85% of samples were classified correctly. Gene set analysis with the false discovery rate set at 0.25 identified several pathways that were differentially regulated when comparing each gastric cancer subtype to adjacent normal stomach. Conclusions: Subtypes of gastric cancer that have epidemiologic and histologic distinctions are also distinguished by gene expression data. These preliminary data suggest a new classification of gastric cancer with implications for improving our understanding of disease biology and identification of unique molecular drivers for each gastric cancer subtype. Clin Cancer Res; 17(9); 2693–701. ©2011 AACR.

Previous studies show that LDH-A knockdown reduces orthotopic 4T1 breast tumor lactate and delays tumor growth and the development of metastases in nude mice. Here, we report significant changes in the tumor microenvironment (TME) and a more robust anti-tumor response in immune competent BALB/c mice. 4T1 murine breast cancer cells were transfected with shRNA plasmids directed against LDH-A (KD) or a scrambled control plasmid (NC). Cells were also transduced with dual luciferase-based reporter systems to monitor HIF-1 activity and the development of metastases by bioluminescence imaging, using HRE-sensitive and constitutive promoters, respectively. The growth and metastatic profile of orthotopic 4T1 tumors developed from these cell lines were compared and a primary tumor resection model was studied to simulate the clinical management of breast cancer. Primary tumor growth, metastasis formation and TME phenotype were significantly different in LDH-A KD tumors compared with controls. In LDH-A KD cells, HIF-1 activity, hexokinase 1 and 2 expression and VEGF secretion were reduced. Differences in the TME included lower HIF-1α expression that correlated with lower vascularity and pimonidazole staining, higher infiltration of CD3+ and CD4+ T cells and less infiltration of TAMs. These changes resulted in a greater delay in metastases formation and 40% long-term survivors (>20 weeks) in the LDH-A KD cohort following surgical resection of the primary tumor. We show for the first time that LDH-depletion inhibits the formation of metastases and prolongs survival of mice through changes in tumor microenvironment that modulate the immune response. We attribute these effects to diminished HIF-1 activity, vascularization, necrosis formation and immune suppression in immune competent animals. Gene-expression analyses from four human breast cancer datasets are consistent with these results, and further demonstrate the link between glycolysis and immune suppression in breast cancer.

Objective— MicroRNAs (miRNAs) are small noncoding RNAs that have the capacity to control protein production through binding “seed” sequences within a target mRNA. Each miRNA is capable of potentially controlling hundreds of genes. The regulation of miRNAs in the lung during the development of pulmonary arterial hypertension (PAH) is unknown. Methods and Results— We screened lung miRNA profiles in a longitudinal and crossover design during the development of PAH caused by chronic hypoxia or monocrotaline in rats. We identified reduced expression of Dicer, involved in miRNA processing, during the onset of PAH after hypoxia. MiR-22, miR-30, and let-7f were downregulated, whereas miR-322 and miR-451 were upregulated significantly during the development of PAH in both models. Differences were observed between monocrotaline and chronic hypoxia. For example, miR-21 and let-7a were significantly reduced only in monocrotaline-treated rats. MiRNAs that were significantly regulated were validated by quantitative polymerase chain reaction. By using in vitro studies, we demonstrated that hypoxia and growth factors implicated in PAH induced similar changes in miRNA expression. Furthermore, we confirmed miR-21 downregulation in human lung tissue and serum from patients with idiopathic PAH. Conclusion— Defined miRNAs are regulated during the development of PAH in rats. Therefore, miRNAs may contribute to the pathogenesis of PAH and represent a novel opportunity for therapeutic intervention.

BACKGROUND Uveal melanoma is the most common intraocular cancer. There are no effective therapies for metastatic disease. Mutations in GNAQ, the gene encoding an alpha subunit of heterotrimeric G proteins, are found in 40% of uveal melanomas. METHODS We sequenced exon 5 of GNAQ and GNA11, a paralogue of GNAQ, in 713 melanocytic neoplasms of different types (186 uveal melanomas, 139 blue nevi, 106 other nevi, and 282 other melanomas). We sequenced exon 4 of GNAQ and GNA11 in 453 of these samples and in all coding exons of GNAQ and GNA11 in 97 uveal melanomas and 45 blue nevi. RESULTS We found somatic mutations in exon 5 (affecting Q209) and in exon 4 (affecting R183) in both GNA11 and GNAQ, in a mutually exclusive pattern. Mutations affecting Q209 in GNA11 were present in 7% of blue nevi, 32% of primary uveal melanomas, and 57% of uveal melanoma metastases. In contrast, we observed Q209 mutations in GNAQ in 55% of blue nevi, 45% of uveal melanomas, and 22% of uveal melanoma metastases. Mutations affecting R183 in either GNAQ or GNA11 were less prevalent (2% of blue nevi and 6% of uveal melanomas) than the Q209 mutations. Mutations in GNA11 induced spontaneously metastasizing tumors in a mouse model and activated the mitogen-activated protein kinase pathway. CONCLUSIONS Of the uveal melanomas we analyzed, 83% had somatic mutations in GNAQ or GNA11. Constitutive activation of the pathway involving these two genes appears to be a major contributor to the development of uveal melanoma. (Funded by the National Institutes of Health and others.).

Animal microRNAs (miRNAs) regulate gene expression by inhibiting translation and/or by inducing degradation of target messenger RNAs. It is unknown how much translational control is exerted by miRNAs on a genome-wide scale. We used a new proteomic approach to measure changes in synthesis of several thousand proteins in response to miRNA transfection or endogenous miRNA knockdown. In parallel, we quantified mRNA levels using microarrays. Here we show that a single miRNA can repress the production of hundreds of proteins, but that this repression is typically relatively mild. A number of known features of the miRNA-binding site such as the seed sequence also govern repression of human protein synthesis, and we report additional target sequence characteristics. We demonstrate that, in addition to downregulating mRNA levels, miRNAs also directly repress translation of hundreds of genes. Finally, our data suggest that a miRNA can, by direct or indirect effects, tune protein synthesis from thousands of genes.

We previously published a report on the comprehensive evaluation of RNA-seq differential analysis (DE) methods [1] where we compared a number of popular DE tools using a variety of different criteria. Since publication we received valuable feedback and suggestions from the community including from the authors of the algorithms. Here we report on two errors that came to our attention following publication. 1. Soon after publication we were notified about a discrepancy in table 2. The correction was posted as comment to the main article and now included as erratum. In the last row of table 2 called "Runtime for experiments with 3-5 replicates…" the values for edgeR and limmaVoom should be Seconds not Minutes. For completeness, the following table is the runtime performance (in seconds) of six DE analysis methods, comparing the 5 replicates from groups A and B from the SEQC data, as measured on Red Hat Enterprise Linux Server release 5.4, with 12 dual cores Intel Xeon 3.33GHz, and 100G RAM. 2. Zhou and Robinson performed a follow up analysis to our manuscript where they reanalyzed DE when genes are expressed in one condition (see correspondence to Rapaport et al.). They identified a coding error in the calculation of edgeR signal to noise values due to incorrect normalization of edgeR count values. Library size values were not used to scale gene counts. This coding error was fixed and a corrected version is deposited in the source code repository available at: http://bitbucket.org/soccin/seqc. We redid the analysis for evaluation of genes expressed in only one condition and the main conclusions remain unchanged. The corrected version of Figure 4 in the main manuscript is presented here as Fig. ​Fig.11. Fig. 1 a edgeR correlation between signal-to-noise and –log10(p-values) using the corrected normalization of gene counts. b ROC analysis of curves for detection of DE at signal-to-noise ratio of ≥3. Note that edgeR AUC has improved from 0.788 ... We note that our reanalysis is based on the exact same procedure used in the original publication. However, Zhou and Robinson introduce a different version of this analysis where the ROC analysis is based on a common set of genes. Second, most packages used in our original publication have since been updated and in some cases the algorithms revised substantially. Therefore, it is possible that a similar comparison with the latest versions of the packages may result in different conclusions.

The choice of algorithmic tool for analysis of genomic data is often guided by personal preference and convenience rather than clear scientific criteria. In a recent paper (Rapaport et al. Genome Biol. 2013;14:R95) we performed an objective analysis of widely used RNA-seq differential expression (DE) programs in order to understand the performance of the various tools with respect to different evaluation measures. One such measure was the interpretation of DE when there is no detectable signal from one of the contrasting conditions. Zhou and Robinson present insightful reanalysis of this comparison. They present two claims. The first is that there was a coding error when calculating the normalized edgeR count values. Second, that our use of signal-to-noise ratio (S/N) as a measure of sensitivity and specificity (in the form of ROC) is incorrect. This is because the mean and variance are correlated in count data and therefore the signal (mean) and noise (standard deviation) are not two independent measures. Perhaps, more importantly, the ROC analysis was not based on the same set of truth set for all methods since the same S/N threshold value corresponds to very different genes sets for each method. The authors are correct in both these points and we appreciate the effort invested in bringing those to light. Our coding error was omitting scaling by library size when calculating the normalized edgeR values. However, this error has very little impact on the analysis as presented in Fig. 4 of our manuscript and does not change its conclusions(see erratum to the manuscript). The most salient point raised is the use of the same S/N threshold in the ROC evaluation. The main argument is that the scale of mean and variance is different between the different methods and therefore the methods are not evaluating the same sets of genes. That is correct and we agree with these points. The second point raised is that since the mean and variance are highly correlated in count data calculating S/N using transformed mean variance values is more appropriate (Fig. 2 in Zhou and Robinson’s response). It is notable that the ROC performances of edgeR and DESeq2 relative to limma in Fig. 2d are similar to those presented in Fig. 4b of Rapaport et al. suggesting that selecting a fixed set of labels is not the major reason for the differences in performances. The improvement in the ROC performance for edgeR (and DESeq2) relative to limma is attributed to the transformation of the mean variance values (Fig. 2b and c). The ROC analysis is only one part of much broader analyses of this subset of genes. Our analysis was originally motivated by the observation that often genes with similar mean expression values, in the condition in which they are expressed, had different P values. This was contrary to our initial expectation, since these genes had similar mean expression values and therefore similar log expression ratios we anticipated similar P values. Moreover, we noticed that some genes with higher expression values had less significant P values than genes with lower expression values (and similar standard deviation). This introduces a problem when selecting genes for subsequent expression validation and functional importance. We were interested in characterizing this further and gain a better understanding of this discrepancy. As noted in both Rapaport et al. and in the response from Zhou and Robinson, most of these genes are clearly differentially expressed and are correctly detected as such by all methods. Hence the challenge is not in correct identification of DE (that is, rejecting the null hypothesis) but rather meaningful ordering of the DE genes. Therefore, ranking this specific set of genes by their P values of differential expression indicates the confidence in experimental measurement rather than rejection of the null hypothesis (which is trivial in this case since there is a clear change in expression). In our paper (Fig. 4a) we show that limma and PoissonSeq packages have a clear monotonic relationship between the S/N and P value assigned to difference in expression for this particular subset of genes with zero value in one condition. In contrast, DESeq, edgeR, and cuffdiff did not display this monotonic relationship. It would be interesting to see the impact of mean variance scaling on this correlation.

The basic underlying problem in reverse engineering of gene regulatory networks from gene expression data is that the expression of a gene encoding the regulator provides only limited information about its protein activity. The proteins, which result from translation, are subject to stringent posttranscriptional control and modification. Often, it is only the modified version of the protein that is capable of activating or repressing its regulatory targets. At present there exists no reliable high-throughput technology to measure the protein activity levels in real-time, and therefore they are, so-to-say, lost in translation. However, these activity levels can be recovered by studying the gene expression of their targets. Here, we describe a computational approach to predict temporal regulator activity levels from the gene expression of its transcriptional targets in a network motif with one regulator and many targets. We consider an example of an SOS repair system, and computationally infer the regulator activity of its master repressor, LexA. The reconstructed activity profile of LexA exhibits a behavior that is similar to the experimentally measured profile of this repressor: after UV irradiation, the amount of LexA substantially decreases within a few minutes, followed by a recovery to its normal level. Our approach can easily be applied to known single-input motifs in other organisms.

This review describes the development of the molecular level Ca(2+)-voltage hypothesis. Theoretical considerations and feedback between theory and experiments played a key role in its development. The theory, backed by experiments, states that at fast synapses, membrane potential by means of presynaptic inhibitory autoreceptors controls initiation and termination of neurotransmitter release. A molecular kinetic scheme which depicts initiation and termination of evoked release is discussed. This scheme is able to account for both spontaneous release and evoked release. The physiological implications of this scheme are enumerated.

Summary Much biological and medical research employs microarray studies to monitor gene expression levels across a wide range of organisms and under many experimental conditions. Dual channel microarrays are a common platform and allow two samples to be measured simultaneously. A frequently used design uses a common reference sample to make conditions across different arrays comparable. Our aim is to formulate microarray experiments in the experimental design context and to use simulated annealing to search for near-optimal designs. We identify a subclass of designs, the so-called interwoven loop designs, that seems to have good optimality properties compared with the near-optimal designs that are found by simulated annealing. Commonly used reference designs and dye swap designs are shown to be highly inefficient.

The topic of multibody analysis deals with the automatic generation and subsequent solution of the equations of motion for a system of interconnected bodies. Many academic and industrial computer programs have been developed to carry out this task. However, although some of these programs can obtain the equations of motion in fully symbolic form, it is believed that all the existing programs for multibody analysis solve these equations numerically. The idea behind the research which underpins this paper is to select and implement a symbolic version of an existing multibody algorithm and then to integrate it with a recently developed solver which can obtain approximate analytical solutions to the equations of motion. The present paper deals with the selection of this multibody method. The theory behind the chosen method (the Roberson and Schwertassek algorithm) is described in some detail but also in a much more concise form than can be found in the literature. In addition, a fully worked example of the application of the multibody algorithm to a practical physical problem is given. Such examples are rare in the literature, and so it is intended that this paper can serve as a basis for enhanced didactic practice in this traditionally difficult subject area.

Motivation: Despite theoretical arguments that so-called ‘loop designs’ for two-channel DNA microarray experiments are more efficient, biologists continue to use ‘reference designs’. We describe two sets of microarray experiments with RNA from two different biological systems (TPA-stimulated mammalian cells and Streptomyces coelicolor). In each case, both a loop and a reference design were used with the same RNA preparations with the aim of studying their relative efficiency. Results: The results of these experiments show that (1) the loop design attains a much higher precision than the reference design, (2) multiplicative spot effects are a large source of variability, and if they are not accounted for in the mathematical model, for example, by taking log-ratios or including spot effects, then the model will perform poorly. The first result is reinforced by a simulation study. Practical recommendations are given on how simple loop designs can be extended to more realistic experimental designs and how standard statistical methods allow the experimentalist to use and interpret the results from loop designs in practice. Availability: The data and R code are available at http://exgen.ma.umist.ac.uk Contact: veronica.vinciotti@brunel.ac.uk

In this study, the development of a symbolic multibody code is described. This code uses Lagrange's equations to derive automatically the state-space equations of motion, in relative coordinates, for open-loop systems made up of rigid bodies and flexible Timoshenko beams. The algorithm is then encoded in Mathematica as the package MultiFlex.m and the straightforward application of MultiFlex to two examples with 1 degree of freedom and 14 degrees of freedom, respectively, is presented. The MultiFlex package represents one part of a suite of programs which are being developed by the authors in order to provide entirely symbolic analysis of multi-degree of freedom problems in engineering dynamics.

The paper considers an extension of the Multiple Scales method to strongly non-linear vibration systems. It is well known that similarly to other perturbation methods, the method of Multiple Scales produces good accuracy results when the perturbation parameter is small which is not the case in many practical applications. A two-step hybrid technique aimed at dealing with strong nonlinearities has originally been developed by Noor [9]. It is based on combining a perturbation method with a direct variational procedure. The full potential of such a hybrid technique has not yet been fully realized in solving non-linear vibrational problems. This paper examines the hybrid technique of combining the Multiple Scale method with the Galerkin procedure. Firstly, a perturbation solution is generated assuming a perturbation parameter is small. The next step involves using the computed perturbation functions as the coordinate (or approximation) modes whose amplitudes are computed by applying Bubnov-Galerkin conditions. The effectiveness of the hybrid Multiple Scales-Galerkin procedure is demonstrated on a 2-degree-of-freedom autoparametric vibration absorber by comparing the solutions computed by the method of Multiple Scales and by the hybrid method with the numerical results.

This paper considers an accepted model of the metal cutting process dynamics in the context of an approximate analysis of the resulting non-linear differential equations of motion. The process model is based upon the established mechanics of orthogonal cutting and results in a pair of non-linear ordinary differential equations which are then restated in a form suitable for approximate analytical solution. The chosen solution technique is the perturbation method of multiple time scales and approximate closed-form solutions are generated for the most important non-resonant case. Numerical data are then substituted into the analytical solutions and key results are obtained and presented. Some comparisons between the exact numerical calculations for the forces involved and their reduced and simplified analytical counterparts are given. It is shown that there is almost no discernible difference between the two thus confirming the validity of the excitation functions adopted in the analysis for the data sets used, these being chosen to represent a real orthogonal cutting process. In an attempt to provide guidance for the selection of technological parameters for the avoidance of primary chatter, this paper determines for the first time the stability regions in terms of the depth of cut and the cutting speed co-ordinates.

The last decade has seen a marked rise in the use of cancer tissues obtained from research autopsies. Such resources have been invaluable for studying cancer evolution or the mechanisms of therapeutic resistance to targeted therapies. Degradation of biomolecules is a potential challenge to usage of cancer tissues obtained in the post-mortem setting and remains incompletely studied. We analysed the nucleic acid quality in 371 different frozen tissue samples collected from 80 patients who underwent a research autopsy, including eight normal tissue types, primary and metastatic tumors. Our results indicate that RNA integrity number (RIN) of normal tissues decline with the elongation of post-mortem interval (PMI) in a tissue-type specific manner. Unlike normal tissues, the RNA quality of cancer tissues is highly variable with respect to post-mortem interval. The kinetics of DNA damage also has tissue type-specific features. Moreover, while DNA degradation is an indicator of low RNA quality, the converse is not true. Finally, we show that despite RIN values as low as 5.0, robust data can be obtained by RNA sequencing that reliably discriminates expression signatures. Based on this promising pilot data, we are currently identifying samples for multiregion RNA sequencing of pancreatic cancers with divergent morphologies within the primary and/or metastatic sites. Thus far we have screened 554 RNA samples from 10 autopsy cases. These samples represent 198 geographically distinct regions of tumor with duplicate or triplicate RNA extractions. The number of samples per patient ranges from 17 to 86 with an average of 55. Our data shows that high quality RNA can be acquired from cases with long PMI (>48h). While extensive screening of samples is required to identify those with optimal quality, the downstream data produced is of high quality and can support novel hypotheses or validations of experimental data. Citation Format: Jun Fan, Raya Khanin, Hitomi Sakamoto, Yi Zhong, Chelsea Michael, Derwin Pena, Breanna Javier, Laura Wood, Christine Iacobuzio-Donahue. Quantification of nucleic acid quality in postmortem tissues from a cancer research autopsy program [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 504. doi:10.1158/1538-7445.AM2017-504

There are conflicting reports on the impact of soy on breast carcinogenesis. This study examines the effects of soy supplementation on breast cancer-related genes and pathways.Women (n = 140) with early-stage breast cancer were randomly assigned to soy protein supplementation (n = 70) or placebo (n = 70) for 7 to 30 days, from diagnosis until surgery. Adherence was determined by plasma isoflavones: genistein and daidzein. Gene expression changes were evaluated by NanoString in pre- and posttreatment tumor tissue. Genome-wide expression analysis was performed on posttreatment tissue. Proliferation (Ki67) and apoptosis (Cas3) were assessed by immunohistochemistry.Plasma isoflavones rose in the soy group (two-sided Wilcoxon rank-sum test, P.001) and did not change in the placebo group. In paired analysis of pre- and posttreatment samples, 21 genes (out of 202) showed altered expression (two-sided Student's t-test, P.05). Several genes including FANCC and UGT2A1 revealed different magnitude and direction of expression changes between the two groups (two-sided Student's t-test, P.05). A high-genistein signature consisting of 126 differentially expressed genes was identified from microarray analysis of tumors. This signature was characterized by overexpression (2-fold) of cell cycle transcripts, including those that promote cell proliferation, such as FGFR2, E2F5, BUB1, CCNB2, MYBL2, CDK1, and CDC20 (P.01). Soy intake did not result in statistically significant changes in Ki67 or Cas3.Gene expression associated with soy intake and high plasma genistein defines a signature characterized by overexpression of FGFR2 and genes that drive cell cycle and proliferation pathways. These findings raise the concerns that in a subset of women soy could adversely affect gene expression in breast cancer.