Explore the vast range of titles available.

Convolutional neural networks (CNNs) have been successfully used in many applications where important information about data is embedded in the order of features, such as speech and imaging. However, most tabular data do not assume a spatial relationship between features, and thus are unsuitable for modeling using CNNs. To meet this challenge, we develop a novel algorithm, image generator for tabular data (IGTD), to transform tabular data into images by assigning features to pixel positions so that similar features are close to each other in the image. The algorithm searches for an optimized assignment by minimizing the difference between the ranking of distances between features and the ranking of distances between their assigned pixels in the image. We apply IGTD to transform gene expression profiles of cancer cell lines (CCLs) and molecular descriptors of drugs into their respective image representations. Compared with existing transformation methods, IGTD generates compact image representations with better preservation of feature neighborhood structure. Evaluated on benchmark drug screening datasets, CNNs trained on IGTD image representations of CCLs and drugs exhibit a better performance of predicting anti-cancer drug response than both CNNs trained on alternative image representations and prediction models trained on the original tabular data.

Background In order to correctly decode phenotypic information from RNA-sequencing (RNA-seq) data, careful selection of the RNA-seq quantification measure is critical for inter-sample comparisons and for downstream analyses, such as differential gene expression between two or more conditions. Several methods have been proposed and continue to be used. However, a consensus has not been reached regarding the best gene expression quantification method for RNA-seq data analysis. Methods In the present study, we used replicate samples from each of 20 patient-derived xenograft (PDX) models spanning 15 tumor types, for a total of 61 human tumor xenograft samples available through the NCI patient-derived model repository (PDMR). We compared the reproducibility across replicate samples based on TPM (transcripts per million), FPKM (fragments per kilobase of transcript per million fragments mapped), and normalized counts using coefficient of variation, intraclass correlation coefficient, and cluster analysis. Results Our results revealed that hierarchical clustering on normalized count data tended to group replicate samples from the same PDX model together more accurately than TPM and FPKM data. Furthermore, normalized count data were observed to have the lowest median coefficient of variation (CV), and highest intraclass correlation (ICC) values across all replicate samples from the same model and for the same gene across all PDX models compared to TPM and FPKM data. Conclusion We provided compelling evidence for a preferred quantification measure to conduct downstream analyses of PDX RNA-seq data. To our knowledge, this is the first comparative study of RNA-seq data quantification measures conducted on PDX models, which are known to be inherently more variable than cell line models. Our findings are consistent with what others have shown for human tumors and cell lines and add further support to the thesis that normalized counts are the best choice for the analysis of RNA-seq data across samples.

Role of DNA damage and demethylation on anticancer activity of DNA methyltransferase inhibitors (DNMTi) remains undefined. We report the effects of DNMT1 gene deletion/disruption ( DNMT1 −/− ) on anticancer activity of a class of DNMTi in vitro, in vivo and in human cancers. The gene deletion markedly attenuated cytotoxicity and growth inhibition mediated by decitabine, azacitidine and 5-aza-4′-thio-2′-deoxycytidine (aza-T-dCyd) in colon and breast cancer cells. The drugs induced DNA damage that concurred with DNMT1 inhibition, subsequent G 2 /M cell-cycle arrest and apoptosis, and upregulated p21 in DNMT1 + / +  versus DNMT1 −/− status, with aza-T-dCyd the most potent. Tumor growth and DNMT1 were significantly inhibited, and p21 was upmodulated in mice bearing HCT116 DNMT1 +/+  xenograft and bladder PDX tumors. DNMT1 gene deletion occurred in ~ 9% human colon cancers and other cancer types at varying degrees. Decitabine and azacitidine demethylated CDKN2A / CDKN2B genes in DNMT1 + / +  and DNMT1 −/− conditions and increased histone-H3 acetylation with re-expression of p16 INK4A /p15 INK4B in DNMT1 −/− state. Thus, DNMT1 deletion confers resistance to DNMTi, and their anti-cancer activity is determined by DNA damage effects. Patients with DNMT1 gene deletions may not respond to DNMTi treatment.

Background The National Cancer Institute drug pair screening effort against 60 well-characterized human tumor cell lines (NCI-60) presents an unprecedented resource for modeling combinational drug activity. Results We present a computational model for predicting cell line response to a subset of drug pairs in the NCI-ALMANAC database. Based on residual neural networks for encoding features as well as predicting tumor growth, our model explains 94% of the response variance. While our best result is achieved with a combination of molecular feature types (gene expression, microRNA and proteome), we show that most of the predictive power comes from drug descriptors. To further demonstrate value in detecting anticancer therapy, we rank the drug pairs for each cell line based on model predicted combination effect and recover 80% of the top pairs with enhanced activity. Conclusions We present promising results in applying deep learning to predicting combinational drug response. Our feature analysis indicates screening data involving more cell lines are needed for the models to make better use of molecular features.

DNA-damaging agents (DDAs) constitute the backbone of treatment for most human tumors. Here we used the National Cancer Institute Antitumor Cell Line Panel (the NCI-60) to identify predictors of cancer cell response to topoisomerase I (Top1) inhibitors, a widely used class of DDAs. We assessed the NCI-60 transcriptome using Affymetrix Human Exon 1.0 ST microarrays and correlated the in vitro activity of four Top1 inhibitors with gene expression in the 60 cell lines. A single gene, Schlafen-11 (SLFN11), showed an extremely significant positive correlation with the response not only to Top1 inhibitors, but also to Top2 inhibitors, alkylating agents, and DNA synthesis inhibitors. Using cells with endogenously high and low SLFN11 expression and siRNA-mediated silencing, we show that SLFN11 is causative in determining cell death and cell cycle arrest in response to DDAs in cancer cells from different tissues of origin. We next analyzed SLFN11 expression in ovarian and colorectal cancers and normal corresponding tissues from The Cancer Genome Atlas database and observed that SLFN11 has a wide expression range. We also observed that high SLFN11 expression independently predicts overall survival in a group of ovarian cancer patients treated with cisplatin-containing regimens. We conclude that SLFN11 expression is causally associated with the activity of DDAs in cancer cells, has a broad expression range in colon and ovarian adenocarcinomas, and may behave as a biomarker for prediction of response to DDAs in the clinical setting.

Phosphorylated H2AX (γ-H2AX) is a sensitive marker for DNA double-strand breaks (DSBs), but the variability of H2AX expression in different cell and tissue types makes it difficult to interpret the meaning of the γ-H2AX level. Furthermore, the assays commonly used for γ-H2AX detection utilize laborious and low-throughput microscopy-based methods. We describe here an ELISA assay that measures both phosphorylated H2AX and total H2AX absolute amounts to determine the percentage of γ-H2AX, providing a normalized value representative of the amount of DNA damage. We demonstrate the utility of the assay to measure DSBs introduced by either ionizing radiation or DNA-damaging agents in cultured cells and in xenograft models. Furthermore, utilizing the NCI-60 cancer cell line panel, we show a correlation between the basal fraction of γ-H2AX and cellular mutation levels. This additional application highlights the ability of the assay to measure γ-H2AX levels in many extracts at once, making it possible to correlate findings with other cellular characteristics. Overall, the γ-H2AX ELISA represents a novel approach to quantifying DNA damage, which may lead to a better understanding of mutagenic pathways in cancer and provide a useful biomarker for monitoring the effectiveness of DNA-damaging anticancer agents.

Developing optimal combination strategies for molecularly targeted anticancer drugs is substantially more complex than for traditional chemotherapies. Here, Doroshow and colleagues discuss the lessons learned from the evaluation of combinations of molecularly targeted anticancer agents by the US National Cancer Institute (NCI), and highlight several new approaches that the NCI has initiated to improve the effectiveness of such combinations. The rapid development of new therapeutic agents that target specific molecular pathways involved in tumour cell proliferation provides an unprecedented opportunity to achieve a much higher degree of biochemical specificity than previously possible with traditional chemotherapeutic anticancer agents. However, the lack of specificity of these established chemotherapeutic drugs allowed a relatively straightforward approach to their use in combination therapies. Developing a paradigm for combining new, molecularly targeted agents, on the other hand, is substantially more complex. The abundance of molecular data makes it possible, at least in theory, to predict how such agents might interact across crucial growth control networks. Initial strategies to examine molecularly targeted agent combinations have produced a small number of successes in the clinic. However, for most of these combination strategies, both in preclinical models and in patients, it is not clear whether the agents being combined actually hit their targets to induce growth inhibition. Here, we consider the initial approach of the US National Cancer Institute (NCI) to the evaluation of combinations of molecularly targeted anticancer agents in patients and provide a description of several new approaches that the NCI has initiated to improve the effectiveness of combination-targeted therapy for cancer.

Background Motivated by the size and availability of cell line drug sensitivity data, researchers have been developing machine learning (ML) models for predicting drug response to advance cancer treatment. As drug sensitivity studies continue generating drug response data, a common question is whether the generalization performance of existing prediction models can be further improved with more training data. Methods We utilize empirical learning curves for evaluating and comparing the data scaling properties of two neural networks (NNs) and two gradient boosting decision tree (GBDT) models trained on four cell line drug screening datasets. The learning curves are accurately fitted to a power law model, providing a framework for assessing the data scaling behavior of these models. Results The curves demonstrate that no single model dominates in terms of prediction performance across all datasets and training sizes, thus suggesting that the actual shape of these curves depends on the unique pair of an ML model and a dataset. The multi-input NN (mNN), in which gene expressions of cancer cells and molecular drug descriptors are input into separate subnetworks, outperforms a single-input NN (sNN), where the cell and drug features are concatenated for the input layer. In contrast, a GBDT with hyperparameter tuning exhibits superior performance as compared with both NNs at the lower range of training set sizes for two of the tested datasets, whereas the mNN consistently performs better at the higher range of training sizes. Moreover, the trajectory of the curves suggests that increasing the sample size is expected to further improve prediction scores of both NNs. These observations demonstrate the benefit of using learning curves to evaluate prediction models, providing a broader perspective on the overall data scaling characteristics. Conclusions A fitted power law learning curve provides a forward-looking metric for analyzing prediction performance and can serve as a co-design tool to guide experimental biologists and computational scientists in the design of future experiments in prospective research studies.

The poly (ADP-ribose) polymerase (PARP) family of enzymes plays a critical role in the maintenance of DNA integrity as part of the base excision pathway of DNA repair. PARP1 is overexpressed in a variety of cancers, and its expression has been associated with overall prognosis in cancer, especially breast cancer. A series of new therapeutic agents that are potent inhibitors of the PARP1 and PARP2 isoforms have demonstrated important clinical activity in patients with breast or ovarian cancers that are caused by mutations in either the BRCA1 or 2 genes. Results from such studies may define a new therapeutic paradigm, wherein simultaneous loss of the capacity to repair DNA damage may have antitumor activity in itself, as well as enhance the antineoplastic potential of cytotoxic chemotherapeutic agents.

Ferroptosis is a form of iron-dependent cell death of interest for the development of novel anti-cancer therapies. Ferroptosis research uses a process of elimination based on assumed ferroptosis-specific inducers and inhibitors; these molecules however have off-target effects and cannot provide a comprehensive picture of overlapping pathways. We investigated whether pyroptosis—a form of inflammatory cell death—is initiated in cancer cells following treatment with the ferroptosis inducer RSL3. We treated 6 cancer cell lines with RSL3 alone or in combination with inhibitors of ferroptosis (Ferrostatin-1), caspases (zVADfmk), necroptosis (Necrostatin-1), BID (BI-6C9), or STING (H-151). Biomarkers of pyroptosis and ferroptosis were assessed using our novel quantitative multiplex immunoassay. Increased secretion of pyroptosis-associated cytokines (IL-1α, IL-1β, IL-18), and gasdermin D and E (GSDMD/E) cleavage with parallel loss of respective full-length proteins—both hallmarks of pyroptosis—were recorded in 5/6 cell lines following RSL3 treatment. RSL3 cytotoxicity was blocked by Ferostatin-1; BID and STING inhibitors also prevented GSDMD/E cleavage. We conclude that the ferroptosis-inducer RSL3 triggers pyroptosis in cancer cells; further work is required to elucidate the role of mitochondria in this process. Measurement of pathway-specific protein biomarkers is therefore necessary to identify the exact mechanism of action of novel cytotoxic agents.