Retraction: Genomic signatures to guide the use of chemotherapeutics

The five pillars of computational reproducibility: bioinformatics and beyond

Computational reproducibility is a simple premise in theory, but is difficult to achieve in practice. Building upon past efforts and proposals to maximize reproducibility and rigor in bioinformatics, we present a framework called the five pillars of reproducible computational research. These include (1) literate programming, (2) code version control and sharing, (3) compute environment control, (4) persistent data sharing and (5) documentation. These practices will ensure that computational research work can be reproduced quickly and easily, long into the future. This guide is designed for bioinformatics data analysts and bioinformaticians in training, but should be relevant to other domains of study.

Jian-Pi-Bu-Xue-Formula Alleviates Cyclophosphamide-Induced Myelosuppression via Up-Regulating NRF2/HO1/NQO1 Signaling

Jian-pi-bu-xue-formula (JPBXF), a TCM formula composed of twelve Chinese medicinal herbs, has been used in clinic to ease patients’ state of weakness and fatigue especially after receiving anti-tumor chemotherapy in China. The lack of the phytochemical characterization, detail therapeutic evaluation and mechanism of JPBXF remains the main limitation for its spreading. In this study, we systematically evaluated the effectiveness and underline mechanism of JPBXF on cyclophosphamide (CTX)-induced myelosuppression and identified the main constituents of JPBXF aqueous extract. JPBXF treatments reversed CTX-induced myelosuppression through increasing the number of haematopoietic stem cells (HSCs) and expression of C-kit in bone marrow cells. Simultaneously, JPBXF treatments alleviated CTX-induced blood cells reduction by increasing numbers of RBCs and WBCs and levels of GM-CSF, TPO and EPO in plasma. JPBXF treatments reduced CTX-induced immunosuppression by increasing expressions of CD3, CD4, and CD8a in PBMCs, and recovering structure damages of thymus and spleen. Moreover, JPBXF notably increased the expression of NRF2 compared with CTX group, and subsequently up-regulated HO1 and NQO1 both in mRNA and protein levels. In addition, eighteen compounds were recognized from JPBXF aqueous extract and the potential targets of the identified compounds were predicted. Overall, JPBXF can greatly reverse CTX-induced myelosuppression in C57BL/6 mice, especially in improving the blood and immune function through activating NRF2/HO1/NQO1 signaling pathway, which provides a reliable reference for JPBXF application in clinical. By recognizing eighteen compounds in JPBXF aqueous extract and predicting the underline mechanisms of the identified compounds, our study would provide theoretical guidance for further research of JPBXF.

Statistical significance: p value, 0.05 threshold, and applications to radiomics-reasons for a conservative approach

Here, we summarise the unresolved debate about p value and its dichotomisation. We present the statement of the American Statistical Association against the misuse of statistical significance as well as the proposals to abandon the use of p value and to reduce the significance threshold from 0.05 to 0.005. We highlight reasons for a conservative approach, as clinical research needs dichotomic answers to guide decision-making, in particular in the case of diagnostic imaging and interventional radiology. With a reduced p value threshold, the cost of research could increase while spontaneous research could be reduced. Secondary evidence from systematic reviews/meta-analyses, data sharing, and cost-effective analyses are better ways to mitigate the false discovery rate and lack of reproducibility associated with the use of the 0.05 threshold. Importantly, when reporting p values, authors should always provide the actual value, not only statements of "p < 0.05" or "p ≥ 0.05", because p values give a measure of the degree of data compatibility with the null hypothesis. Notably, radiomics and big data, fuelled by the application of artificial intelligence, involve hundreds/thousands of tested features similarly to other "omics" such as genomics, where a reduction in the significance threshold, based on well-known corrections for multiple testing, has been already adopted.

Ensuring Quality Standards and Reproducible Research for Data Analysis Services in Oncology: A Cooperative Service Model

Modern molecular high-throughput devices, e.g., next-generation sequencing, have transformed medical research. Resulting data sets are usually high-dimensional on a genomic-scale providing multi-factorial information from intertwined molecular and cellular activities of genes and their products. This genomics-revolution installed precision medicine offering breathtaking opportunities for patient's diagnosis and treatment. However, due to the speed of these developments the quality standards of the involved data analyses are lacking behind, as exemplified by the infamous Duke Saga. In this paper, we argue in favor of a two-stage cooperative serve model that couples data generation and data analysis in the most beneficial way from the perspective of a patient to ensure data analysis quality standards including reproducible research.

Diagnostic Biomarkers: Are We Moving from Discovery to Clinical Application?

BACKGROUND

Despite considerable research investment, moving from biomarker discovery to clinical application has presented unique challenges. We aimed to evaluate progress toward clinical application of a sample of molecular- and “omics”-based diagnostic tests over a 10-year period.

METHODS

We used Scopus to locate studies, published before the December 31, 2016, citing 107 original-research articles published in 2006 that assessed the diagnostic value of a molecular- or “omics”-based test. We identified diagnostic studies of the same test and disease and determined whether the article represented progress in the validation of the molecular test. We classified the types of progress: (a) clinical validation (measuring diagnostic accuracy in a series of patients similar to the population in which the test will be used in practice), (b) technical improvement, (c) extended diagnostic application (modification of the diagnostic question attended initially by the test), (d) economic evaluation, or (e) clinical use or implementation.

RESULTS

In the 10-year period analyzed, 4257 articles cited the 107 diagnostic studies; 118 (2.8%) were diagnostic studies of the same test, and of these papers, 25 (21.2%) did not constitute progress toward validation of the test for use in clinical practice (potential research waste). Of the 107 molecular- or “omics”-based tests described in 2006, only 28 (26.2%) appeared to have made progress toward clinical application. Only 4 (9.1%) of 44 proteomics-based tests had made progress toward clinical application.

CONCLUSIONS

Articles evaluating molecular- or “omics”-based diagnostic tests are numerous in biomedical journals. Few tests have made progress toward clinical application in the 10 years following their discovery.

Machine Learning and Radiogenomics: Lessons Learned and Future Directions

Due to the rapid increase in the availability of patient data, there is significant interest in precision medicine that could facilitate the development of a personalized treatment plan for each patient on an individual basis. Radiation oncology is particularly suited for predictive machine learning (ML) models due to the enormous amount of diagnostic data used as input and therapeutic data generated as output. An emerging field in precision radiation oncology that can take advantage of ML approaches is radiogenomics, which is the study of the impact of genomic variations on the sensitivity of normal and tumor tissue to radiation. Currently, patients undergoing radiotherapy are treated using uniform dose constraints specific to the tumor and surrounding normal tissues. This is suboptimal in many ways. First, the dose that can be delivered to the target volume may be insufficient for control but is constrained by the surrounding normal tissue, as dose escalation can lead to significant morbidity and rare. Second, two patients with nearly identical dose distributions can have substantially different acute and late toxicities, resulting in lengthy treatment breaks and suboptimal control, or chronic morbidities leading to poor quality of life. Despite significant advances in radiogenomics, the magnitude of the genetic contribution to radiation response far exceeds our current understanding of individual risk variants. In the field of genomics, ML methods are being used to extract harder-to-detect knowledge, but these methods have yet to fully penetrate radiogenomics. Hence, the goal of this publication is to provide an overview of ML as it applies to radiogenomics. We begin with a brief history of radiogenomics and its relationship to precision medicine. We then introduce ML and compare it to statistical hypothesis testing to reflect on shared lessons and to avoid common pitfalls. Current ML approaches to genome-wide association studies are examined. The application of ML specifically to radiogenomics is next presented. We end with important lessons for the proper integration of ML into radiogenomics.

'Big data' approaches for novel anti-cancer drug discovery

INTRODUCTION

The development of improved cancer therapies is frequently cited as an urgent unmet medical need. Recent advances in platform technologies and the increasing availability of biological 'big data' are providing an unparalleled opportunity to systematically identify the key genes and pathways involved in tumorigenesis. The discoveries made using these new technologies may lead to novel therapeutic interventions. Areas covered: The authors discuss the current approaches that use 'big data' to identify cancer drivers. These approaches include the analysis of genomic sequencing data, pathway data, multi-platform data, identifying genetic interactions such as synthetic lethality and using cell line data. They review how big data is being used to identify novel drug targets. The authors then provide an overview of the available data repositories and tools being used at the forefront of cancer drug discovery. Expert opinion: Targeted therapies based on the genomic events driving the tumour will eventually inform treatment protocols. However, using a tailored approach to treat all tumour patients may require developing a large repertoire of targeted drugs.

Building a Personalized Cancer Treatment System

This paper reports the process by which a personalized cancer treatment system was built, following a user-centered approach. We give some background on personalized cancer treatment, the particular tumor chemosensitivity assay supported by the system, as well as some quality and legal issues related to such health systems. We describe how Contextual Design was applied when building the system. Contextual design is a user-centered design technique involving seven steps. We also provide some details about the system implementation. Finally, we explain how the Think-Aloud protocol and Heuristic Evaluation methods were used to evaluate the system and report its results. A qualitative assessment from the users perspective is also provided. Results from the heuristic evaluation indicate that only one of ten heuristics was missing from the system, while five were partially covered and four were fully covered.

The reproducibility of biomedical research: Sleepers awake!

There is increasing concern about the reliability of biomedical research, with recent articles suggesting that up to 85% of research funding is wasted. This article argues that an important reason for this is the inappropriate use of molecular techniques, particularly in the field of RNA biomarkers, coupled with a tendency to exaggerate the importance of research findings.

Potentially functional SNPs (pfSNPs) as novel genomic predictors of 5-FU response in metastatic colorectal cancer patients

5-Fluorouracil (5-FU) and its pro-drug Capecitabine have been widely used in treating colorectal cancer. However, not all patients will respond to the drug, hence there is a need to develop reliable early predictive biomarkers for 5-FU response. Here, we report a novel potentially functional Single Nucleotide Polymorphism (pfSNP) approach to identify SNPs that may serve as predictive biomarkers of response to 5-FU in Chinese metastatic colorectal cancer (CRC) patients. 1547 pfSNPs and one variable number tandem repeat (VNTR) in 139 genes in 5-FU drug (both PK and PD pathway) and colorectal cancer disease pathways were examined in 2 groups of CRC patients. Shrinkage of liver metastasis measured by RECIST criteria was used as the clinical end point. Four non-responder-specific pfSNPs were found to account for 37.5% of all non-responders (P<0.0003). Five additional pfSNPs were identified from a multivariate model (AUC under ROC = 0.875) that was applied for all other pfSNPs, excluding the non-responder-specific pfSNPs. These pfSNPs, which can differentiate the other non-responders from responders, mainly reside in tumor suppressor genes or genes implicated in colorectal cancer risk. Hence, a total of 9 novel SNPs with potential functional significance may be able to distinguish non-responders from responders to 5-FU. These pfSNPs may be useful biomarkers for predicting response to 5-FU.

miRNA expression patterns in chemoresistant breast cancer tissues

BACKGROUND/AIMS

Breast cancer chemoresistance is a major obstacle to the successful treatment of patients. miRNAs perform critical roles in biological processes, including tumorigenesis and chemoresistance. However, little clinical data are available regarding the relationship between miRNA expression patterns and breast cancer chemoresistance.

METHODS

We created a doxorubicin-resistant MCF-7 (MCF-/Adr) cell line using a pulse-selection method; then verified the resistance of the MCF-7/Adr cell line to doxorubicin by using the methyl thiazolyl tetrazolium (MTT) assay, terminal deoxyribonucleotidyl transferase (TdT)-mediated dUTP nick-end labeling (TUNEL) staining, and Intracellular doxorubicin accumulation assay. Then, we performed qRT-PCR to detect the expression patterns of 14 selected miRNAs (which are related to breast cancer resistance) in both cell lines. Subsequently, we performed a bioinformatics analysis, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses, to determine the putative functions of 13 differentially expressed miRNA-targeted genes. Finally, we tested the expression levels of these 13 miRNAs in 10 chemotherapy non-responder breast cancer tissues and 29 responder tissues. All statistical analyses were performed by a two-tailed Student's t-test, and a P value less than 0.05 was considered statistically significant.

RESULTS

The results of the MTT assay showed that the MCF-7/Adr cell line was significantly more resistant to doxorubicin compared to the MCF-7 cells The results of the TUNEL assay indicated that doxorubicin induced an increase in the number apoptotic cells in the MCF-7 group. Additionally, the accumulation of doxorubicin was higher in MCF-7 cells compared to MCF-7/Adr cells, which was consistent with the MTT and TUNEL results. The qRT-PCR results demonstrated that compared to the parental MCF-7 cell line, miR-200a, miR-141, miR-200c, miR-31, miR-429, and miR-196b were over-expressed, and let-7e, miR-576-3p, miR-125b-1, miR-370, miR-145, miR-765, and miR-760 were significantly down-regulated in MCF-7/Adr cells. The GO analysis results revealed that the predicted target genes of these 14 miRNAs primarily regulated protein binding, zinc ion binding, DNA binding, and transcription factor activity. The KEGG data demonstrated that these target genes are mainly involved in the MAPK signaling pathway, regulation of the actin cytoskeleton, cytokine-cytokine receptor interaction, and other signaling pathways. Compared to the breast cancer tissues from chemotherapy responders, 10 miRNAs were identified to be dysregulated in the chemoresistant breast cancer tissues. Three of these miRNAs were up-regulated (miR-141, miR-200c, and miR-31), and 7 were down-regulated (let-7e, miR-576-3p, miR-125b-1, miR-370, miR-145, miR-765, and miR-760).

CONCLUSION

In this study, we identified 10 dysregulated miRNAs in both breast cancer cells and chemoresistant tissues, which might be biomarkers for the prognosis of breast cancer chemoresistance. Our study contributes to a comprehensive understanding of prognostic biomarkers during clinical treatment, and we hypothesize that the miRNA signatures of drug-resistant carcinoma tissues could be useful for developing new strategies for targeted therapies in patients with chemoresistant breast cancer.

Perspectives on systems biology applications in diabetic kidney disease

Diabetic kidney disease (DKD) is a microvascular complication of type 1 and 2 diabetes with a devastating impact on individuals with the disease, their families, and society as a whole. DKD is the single most frequent cause of incident chronic kidney disease cases and accounts for over 40% of the population with end-stage renal disease. Contributing factors for the high prevalence are the increase in obesity and subsequent diabetes combined with an improved long-term survival with diabetes. Environment and genetic variations contribute to DKD susceptibility and progressive loss of kidney function. How the molecular mechanisms of genetic and environmental exposures interact during DKD initiation and progression is the focus of ongoing research efforts. The development of standardized, unbiased high-throughput profiling technologies of human DKD samples opens new avenues in capturing the multiple layers of DKD pathobiology. These techniques routinely interrogate analytes on a genome-wide scale generating comprehensive DKD-associated fingerprints. Linking the molecular fingerprints to deep clinical phenotypes may ultimately elucidate the intricate molecular interplay in a disease stage and subtype-specific manner. This insight will form the basis for accurate prognosis and facilitate targeted therapeutic interventions. In this review, we present ongoing efforts from large-scale data integration translating "-omics" research efforts into improved and individualized health care in DKD.

Molecular signatures of lung cancer: defining new diagnostic and therapeutic paradigms

Molecular profiling holds great promise for improving our ability to diagnose, prognosticate, and select individualized treatments for lung cancer patients. However, using multidimensional data and novel technologies to derive these profiles is limited by our ability to employ the assay in a clinical scenario where it can impact the course of disease. Although many molecular signatures have been reported in lung cancer, as of yet, few have been sufficiently validated for widespread clinical use. Recently, several novel signatures have been reported, which address critical aspects of patient care and/or demonstrate improved efforts for appropriate clinical validation. Here, we present our opinion on the current state of the field of molecular signatures in lung cancer.

Parallel evolution under chemotherapy pressure in 29 breast cancer cell lines results in dissimilar mechanisms of resistance

Background

Developing chemotherapy resistant cell lines can help to identify markers of resistance. Instead of using a panel of highly heterogeneous cell lines, we assumed that truly robust and convergent pattern of resistance can be identified in multiple parallel engineered derivatives of only a few parental cell lines.

Methods

Parallel cell populations were initiated for two breast cancer cell lines (MDA-MB-231 and MCF-7) and these were treated independently for 18 months with doxorubicin or paclitaxel. IC50 values against 4 chemotherapy agents were determined to measure cross-resistance. Chromosomal instability and karyotypic changes were determined by cytogenetics. TaqMan RT-PCR measurements were performed for resistance-candidate genes. Pgp activity was measured by FACS.

Results

All together 16 doxorubicin- and 13 paclitaxel-treated cell lines were developed showing 2–46 fold and 3–28 fold increase in resistance, respectively. The RT-PCR and FACS analyses confirmed changes in tubulin isofom composition, TOP2A and MVP expression and activity of transport pumps (ABCB1, ABCG2). Cytogenetics showed less chromosomes but more structural aberrations in the resistant cells.

Conclusion

We surpassed previous studies by parallel developing a massive number of cell lines to investigate chemoresistance. While the heterogeneity caused evolution of multiple resistant clones with different resistance characteristics, the activation of only a few mechanisms were sufficient in one cell line to achieve resistance.

Array-based pharmacogenomics of molecular-targeted therapies in oncology

The advent of microarrays over the past decade has transformed the way genome-wide studies are designed and conducted, leading to an unprecedented speed of acquisition and amount of new knowledge. Microarray data have led to the identification of molecular subclasses of solid tumors characterized by distinct oncogenic pathways, as well as the development of multigene prognostic or predictive models equivalent or superior to those of established clinical parameters. In the field of molecular-targeted therapy for cancer, in particular, the application of array-based methodologies has enabled the identification of molecular targets with 'key' roles in neoplastic transformation or tumor progression and the subsequent development of targeted agents, which are most likely to be active in the specific molecular setting. Herein, we present a summary of the main applications of whole-genome expression microarrays in the field of molecular-targeted therapies for solid tumors and we discuss their potential in the clinical setting. An emphasis is given on deciphering the molecular mechanisms of drug action, identifying novel therapeutic targets and suitable agents to target them with, and discovering molecular markers/signatures that predict response to therapy or optimal drug dose for each patient.

Practical ethics: establishing a pathway to benefit for complex pharmacogenomic tests

Pharmacogenomic tests offer a promising strategy to improve the safety and efficacy of drug treatment. Compelling examples, such as HLA-B*5701 testing to identify patients at risk for abacavir-associated hypersensitivity, are already changing clinical care. However, the level of evidence required to establish clinical utility is often the subject of debate. Determining the most efficient and effective pathway to benefit for a given test is therefore both a practical and an ethical concern.

Gene expression profiling of breast tumor cell lines to predict for therapeutic response to microtubule-stabilizing agents

Microtubule-targeting agents, including taxanes (Tax) and ixabepilone (Ixa), are important components of modern breast cancer chemotherapy regimens, but no molecular parameter is currently available that can predict for their efficiency. We sought to develop pharmacogenomic predictors of Tax- and Ixa-response from a large panel of human breast tumor cell lines (BTCL), then to evaluate their performance in clinical samples. Thirty-two BTCL, representative of the molecular diversity of breast cancers (BC), were treated in vitro with Tax (paclitaxel (Pac), docetaxel (Doc)), and ixabepilone (Ixa), then classified as drug-sensitive or resistant according to their 50% inhibitory concentrations (IC50s). Baseline gene expression data were obtained using Affymetrix U133 Plus 2.0 human oligonucleotide microarrays. Gene expression set (GES) predictors of response to taxanes were derived, then tested for validation internally and in publicly available gene expression datasets. In vitro IC50s of Pac and Doc were almost identical, whereas some Tax-resistant BTCL retained sensitivity to Ixa. GES predictors for Tax-sensitivity (333 genes) and Ixa-sensitivity (79 genes) were defined. They displayed a limited number of overlapping genes. Both were validated by leave-n-out cross-validation (n = 4; for overall accuracy (OA), P = 0.028 for Tax, and P = 0.0005 for Ixa). The GES predictor of Tax-sensitivity was tested on publicly available external datasets and significantly predicted Pac-sensitivity in 16 BTCL (P = 0.04 for OA), and pathological complete response to Pac-based neoadjuvant chemotherapy in BC patients (P = 0.0045 for OA). Applying Tax and Ixa-GES to a dataset of clinically annotated early BC patients identified subsets of tumors with potentially distinct phenotypes of drug sensitivity: predicted Ixa-sensitive/Tax-resistant BC were significantly (P < 0.05, Fischer's exact test) more frequently ER/PR-positive, Ki67-negative, and luminal subtype than predicted Ixa-resistant/Tax-sensitive BC. Genomic predictors for Tax- and Ixa-sensitivity can be derived from BTCL and may be helpful for better selecting cytotoxic treatment in BC patients.

Microarray-based oncogenic pathway profiling in advanced serous papillary ovarian carcinoma

INTRODUCTION

The identification of specific targets for treatment of ovarian cancer patients remains a challenge. The objective of this study is the analysis of oncogenic pathways in ovarian cancer and their relation with clinical outcome.

METHODOLOGY

A meta-analysis of 6 gene expression datasets was done for oncogenic pathway activation scores: AKT, β-Catenin, BRCA, E2F1, EGFR, ER, HER2, INFα, INFγ, MYC, p53, p63, PI3K, PR, RAS, SRC, STAT3, TNFα, and TGFβ and VEGF-A. Advanced serous papillary tumours from uniformly treated patients were selected (N = 464) to find differences independent from stage-, histology- and treatment biases. Survival and correlations with documented prognostic signatures (wound healing response signature WHR/genomic grade index GGI/invasiveness gene signature IGS) were analysed.

RESULTS

The GGI, WHR, IGS score were unexpectedly increased in chemosensitive versus chemoresistant patients. PR and RAS activation score were associated with survival outcome (p = 0.002;p = 0.004). Increased activations of β-Catenin (p = 0.0009), E2F1 (p = 0.005), PI3K (p = 0.003) and p63 (p = 0.05) were associated with more favourable clinical outcome and were consistently correlated with three prognostic gene signatures.

CONCLUSIONS

Oncogenic pathway profiling of advanced serous ovarian tumours revealed that increased β-Catenin, E2F1, p63, PI3K, PR and RAS-pathway activation scores were significantly associated with favourable clinical outcome. WHR, GGI and IGS scores were unexpectedly increased in chemosensitive tumours. Earlier studies have shown that WHR, GGI and IGS are strongly associated with proliferation and that high-proliferative ovarian tumours are more chemosensitive. These findings may indicate opposite confounding of prognostic versus predictive factors when studying biomarkers in epithelial ovarian cancer.

Neoadjuvant chemotherapy in ER+ HER2- breast cancer: response prediction based on immunohistochemical and molecular characteristics

A pathological complete remission (pCR) is rarely achieved by neoadjuvant chemotherapy in estrogen receptor-positive (ER+) HER2-negative (HER2-) tumors. Therefore, its use might be questionable in specific groups of this tumor type. To select which patients benefit and which could be spared neoadjuvant chemotherapy, we tested standard pathology and molecular markers in ER+ HER2- breast tumors. Pretreatment biopsies were available from 211 ER+ HER2- tumors, who had been treated with neoadjuvant chemotherapy (adriamycin/cyclophosphamide). mRNA expression data were available for 132 tumors. We determined progesterone receptor expression (PR), endocrine sensitivity, HER2 expression, histology, proliferation, and molecular subtypes. We correlated these data to chemotherapy response using pCR rates and the previously published neoadjuvant response index (NRI). PR-negative tumors (n = 65, 30.8%) and luminal B type tumors (n = 43, 20.4%) responded significantly better to chemotherapy than other tumors. These associations remained significant in multivariate analysis. However, even in the subgroup of patients with the lowest response rate, comprising tumors that had both a positive-PR expression and the luminal A subtype (n = 58, 44%), the majority of the patients had downstaging because of chemotherapy. For histology (lobular vs. ductal), endocrine sensitivity, and proliferation, no associations with chemotherapy response were observed. Gene expression array analysis resulted in 28 significant genes (FDR < 0.1). PR expression and luminal B status are associated with a better response to neoadjuvant chemotherapy. However, both markers had only weak response predictive power, and it was not possible to identify a subgroup with no or only minimal chemotherapy benefit. Therefore, the decision to refrain from neoadjuvant chemotherapy to ER+ HER2- breast tumors should not be based on predictive markers, but exclusively on estimates of prognosis.