scholarly journals kESVR: An Ensemble Model for Drug Response Prediction in Precision Medicine Using Cancer Cell Lines Gene Expression

Genes ◽  
2021 ◽  
Vol 12 (6) ◽  
pp. 844
Author(s):  
Abhishek Majumdar ◽  
Yueze Liu ◽  
Yaoqin Lu ◽  
Shaofeng Wu ◽  
Lijun Cheng
Keyword(s):  
Gene Expression ◽  
Support Vector Regression ◽  
Cell Line ◽  
Cell Lines ◽  
Cancer Cell ◽  
Drug Response ◽  
Prediction Models ◽  
Response Prediction ◽  
Cancer Cell Lines ◽  
Support Vector

Background: Cancer cell lines are frequently used in research as in-vitro tumor models. Genomic data and large-scale drug screening have accelerated the right drug selection for cancer patients. Accuracy in drug response prediction is crucial for success. Due to data-type diversity and big data volume, few methods can integrative and efficiently find the principal low-dimensional manifold of the high-dimensional cancer multi-omics data to predict drug response in precision medicine. Method: A novelty k-means Ensemble Support Vector Regression (kESVR) is developed to predict each drug response values for single patient based on cell-line gene expression data. The kESVR is a blend of supervised and unsupervised learning methods and is entirely data driven. It utilizes embedded clustering (Principal Component Analysis and k-means clustering) and local regression (Support Vector Regression) to predict drug response and obtain the global pattern while overcoming missing data and outliers’ noise. Results: We compared the efficiency and accuracy of kESVR to 4 standard machine learning regression models: (1) simple linear regression, (2) support vector regression (3) random forest (quantile regression forest) and (4) back propagation neural network. Our results, which based on drug response across 610 cancer cells from Cancer Cell Line Encyclopedia (CCLE) and Cancer Therapeutics Response Portal (CTRP v2), proved to have the highest accuracy (smallest mean squared error (MSE) measure). We next compared kESVR with existing 17 drug response prediction models based a varied range of methods such as regression, Bayesian inference, matrix factorization and deep learning. After ranking the 18 models based on their accuracy of prediction, kESVR ranks first (best performing) in majority (74%) of the time. As for the remaining (26%) cases, kESVR still ranked in the top five performing models. Conclusion: In this paper we introduce a novel model (kESVR) for drug response prediction using high dimensional cell-line gene expression data. This model outperforms current existing prediction models in terms of prediction accuracy and speed and overcomes overfitting. This can be used in future to develop a robust drug response prediction system for cancer patients using the cancer cell-lines guidance and multi-omics data.

Deep learning for drug response prediction in cancer

2020 ◽  
Author(s):  
Delora Baptista ◽  
Pedro G Ferreira ◽  
Miguel Rocha
Keyword(s):  
Deep Learning ◽  
Cell Lines ◽  
Cancer Cell ◽  
High Throughput Screening ◽  
Drug Response ◽  
Prediction Models ◽  
Response Prediction ◽  
Cancer Cell Lines ◽  
Cancer Treatments ◽  
Novel Drugs

Abstract Predicting the sensitivity of tumors to specific anti-cancer treatments is a challenge of paramount importance for precision medicine. Machine learning(ML) algorithms can be trained on high-throughput screening data to develop models that are able to predict the response of cancer cell lines and patients to novel drugs or drug combinations. Deep learning (DL) refers to a distinct class of ML algorithms that have achieved top-level performance in a variety of fields, including drug discovery. These types of models have unique characteristics that may make them more suitable for the complex task of modeling drug response based on both biological and chemical data, but the application of DL to drug response prediction has been unexplored until very recently. The few studies that have been published have shown promising results, and the use of DL for drug response prediction is beginning to attract greater interest from researchers in the field. In this article, we critically review recently published studies that have employed DL methods to predict drug response in cancer cell lines. We also provide a brief description of DL and the main types of architectures that have been used in these studies. Additionally, we present a selection of publicly available drug screening data resources that can be used to develop drug response prediction models. Finally, we also address the limitations of these approaches and provide a discussion on possible paths for further improvement. Contact:[email protected]

Development of a Gene Expression Panel, for the Prediction of Protein Abundances in Cancer Cell Lines

2021 ◽  
Vol 16 ◽  
Author(s):  
Gunhee Lee ◽  
Yeun-Jun Chung ◽  
Minho Lee
Keyword(s):  
Gene Expression ◽  
Mrna Expression ◽  
Cell Lines ◽  
Cancer Cell ◽  
Prediction Models ◽  
Cancer Cell Lines ◽  
Support Vector ◽  
Rna Expression ◽  
Expression Levels ◽  
Protein Levels

Background: Due to the ease of quantifying mRNA expression in comparison with that of protein abundances, many studies have utilized it to infer protein product quantification. However, the mRNA expression values for a gene and its protein products are not known to have a strong relationship, because of the complex mechanisms required to regulate the amounts of protein levels, from translation to post-translational modifications. Methods: We have developed, in this study, models to predict protein levels from mRNA expression levels using the transcriptome and reverse phase protein arrays (RPPA)-based on protein levels in pan-cancer cell lines. When predicting the abundance of a protein expression, in addition to using RNA expression of the corresponding gene, we also used RNA expression levels of a particular set of other genes. By applying support vector regression, we have identified a 47-gene expression panel that contributes to the improved performance of the prediction, and its optimal subsets specific to each protein species. Result and Conclusion: Eventually, our final prediction models doubled the number of predictable protein expressions (r > 0.7). Due to the weaknesses of RPPA, our model had some limitations, however, we expect that these prediction models and the panel can be widely used in the future to infer protein abundances.

Determining the Relative Gene Expression Level of Hypoxia Related Genes in Different Cancer Cell Lines

2020 ◽  
Vol 14 (1) ◽  
pp. 52-59
Author(s):  
Laila Baqlouq ◽  
Malek Zihlif ◽  
Hana Hammad ◽  
Tuqa M. Abu Thaib
Keyword(s):  
Gene Expression ◽  
Cancer Cells ◽  
Cell Line ◽  
Cell Lines ◽  
Cancer Cell ◽  
Cancer Cell Lines ◽  
Breast Epithelial Cell ◽  
Pcr Array ◽  
Pancreatic Cancer Cells ◽  
Inducible Genes

Objective: This study aims to identify the changes in expression of hypoxia-inducible genes in seven different cancer cell lines that vary in their oxygen levels in an attempt to identify hypoxia biomarkers that can be targeted in therapy. Profiling of hypoxia inducible-gene expression of these different cancer cell lines can be used as a baseline data for further studies. Methods: Human cancer cell lines obtained from the American Type Culture Collection were used; MCF7 breast cancer cells, PANC1 pancreatic cancer cells, PC-3 prostate cancer cells, SH-SY5Y neuroblastoma brain cancer cells, A549 lung cancer cells, and HEPG2 hepatocellular carcinoma. In addition, we used MCF10A non-tumorigenic human breast epithelial cell line as a normal cell line. The differences in gene expression were examined using real-time PCR array (PAHS-032Z, Human Hypoxia Signaling Pathway PCR Array) and analyzed using the ΔΔCt method. Results: Almost all hypoxia-inducible genes showed a PO2-dependent up- and down-regulated expression. Noticeable gene expression differences were identified. The most important changes occurred in the HIF1α and NF-KB signaling pathways targeted genes and in central carbon metabolism pathway genes such as HKs, PFKL, and solute transporters. Conclusion: This study identified possible hypoxia biomarkers genes such as NF-KB, HIF1α, HK, PFKL, and PIM1 that were expressed in all hypoxic cells. Pleotropic pathways that play a role in inducing hypoxia directly such as HIF1 α and NF-kB pathways were upregulated. In addition, genes expressed only in the severe hypoxic liver and pancreatic cells indicate that severe and intermediate hypoxic cancer cells vary in their gene expression. Gene expression differences between cancer and normal cells showed the shift in gene expression profile to survive and proliferate under hypoxia.

A gene expression profile indicative of early stage HER2 tyrosine kinase inhibitor response.

2013 ◽  
Vol 31 (15_suppl) ◽  
pp. e11536-e11536
Author(s):  
Fiona O'Neill ◽  
Stephen F. Madden ◽  
Martin Clynes ◽  
Padraig Doolan ◽  
John Crown ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Gene Expression ◽  
Tyrosine Kinase ◽  
Cell Line ◽  
Expression Pattern ◽  
Cell Lines ◽  
Cancer Cell ◽  
Breast Cancer Cell ◽  
Cancer Cell Lines ◽  
Breast Cancer Cell Lines

e11536 Background: Lapatinib, afatinib and neratinib are tyrosine kinase inhibitors (TKIs) of HER2 and EGFR growth receptors. A panel of breast cancer cell lines was treated with these agents and gefintib (EGFR inhibitor) and the expression pattern of a specific panel of genes investigated as a potential marker of early drug response. Methods: RNA was extracted from breast cancer cell lines (BT474, SKBR3 and MDAMB453) with differing HER2 expression patterns and sensitivity to lapatinib before and 12hrs after treatment with 1 µM of lapatinib, 150nM of afatinib, 150nM of neratinib or 1µM of gefitinib. Gene expression changes were measured by Taqman RT-PCR and RQ values were determined using the comparative cycle threshold (Ct) method. Results: The expression of RB1CC1, ERBB3, FOXO3a, NR3C1 was directly correlated with the degree of sensitivity of the cell line to lapatinib and was observed to “switch” from up-regulated to down-regulated in the HER2 expressing lapatinib insensitive cell line (MDAMD453). The CCND1 gene (functionally linked to the other four genes) demonstrated an inversely proportional response to drug exposure; showing a trend of strong down-regulation in lapatinib-sensitive lines. A similar expression pattern was observed following the treatment with both neratinib and afatinib. In contrast, gefitinib treatment, resulted in a completely different expression pattern change. Conclusions: In these HER2-expressing cell models, lapatinib, neratinib and afatinib treatment generated a common, characteristic and specific gene expression response, proportionate to the sensitivity of the cell lines to the HER2 inhibitor. Characterisation of changes in these genes shortly after drug treatment may therefore give a valuable predictor of the likely extent and specificity of tumour HER2 inhibitor response in patients, potentially guiding more specific use of these agents.

Clinical drug response prediction from preclinical cancer cell lines by logistic matrix factorization approach

2021 ◽  
Author(s):  
Akram Emdadi ◽  
Changiz Eslahchi
Keyword(s):  
Cell Line ◽  
Cell Lines ◽  
Cancer Cell ◽  
Matrix Factorization ◽  
Drug Response ◽  
Drug Sensitivity ◽  
Cancer Cell Lines ◽  
Tissue Type ◽  
Factorization Approach ◽  
Clinical Drug

Predicting tumor drug response using cancer cell line drug response values for a large number of anti-cancer drugs is a significant challenge in personalized medicine. Predicting patient response to drugs from data obtained from preclinical models is made easier by the availability of different knowledge on cell lines and drugs. This paper proposes the TCLMF method, a predictive model for predicting drug response in tumor samples that was trained on preclinical samples and is based on the logistic matrix factorization approach. The TCLMF model is designed based on gene expression profiles, tissue type information, the chemical structure of drugs and drug sensitivity (IC 50) data from cancer cell lines. We use preclinical data from the Genomics of Drug Sensitivity in Cancer dataset (GDSC) to train the proposed drug response model, which we then use to predict drug sensitivity of samples from the Cancer Genome Atlas (TCGA) dataset. The TCLMF approach focuses on identifying successful features of cell lines and drugs in order to calculate the probability of the tumor samples being sensitive to drugs. The closest cell line neighbours for each tumor sample are calculated using a description of similarity between tumor samples and cell lines in this study. The drug response for a new tumor is then calculated by averaging the low-rank features obtained from its neighboring cell lines. We compare the results of the TCLMF model with the results of the previously proposed methods using two databases and two approaches to test the model’s performance. In the first approach, 12 drugs with enough known clinical drug response, considered in previous methods, are studied. For 7 drugs out of 12, the TCLMF can significantly distinguish between patients that are resistance to these drugs and the patients that are sensitive to them. These approaches are converted to classification models using a threshold in the second approach, and the results are compared. The results demonstrate that the TCLMF method provides accurate predictions across the results of the other algorithms. Finally, we accurately classify tumor tissue type using the latent vectors obtained from TCLMF’s logistic matrix factorization process. These findings demonstrate that the TCLMF approach produces effective latent vectors for tumor samples. The source code of the TCLMF method is available in https://github.com/emdadi/TCLMF.

scholarly journals Investigation of the expression of RIF1 gene on head and neck, pancreatic and brain cancer and cancer stem cells

2016 ◽  
Vol 39 (6) ◽  
pp. 43 ◽  
Author(s):  
Hacer E GursesCila ◽  
Muradiye Acar ◽  
Furkan B Barut ◽  
Mehmet Gunduz ◽  
Reidar Grenman ◽  
...  
Keyword(s):  
Gene Expression ◽  
Pancreatic Cancer ◽  
Stem Cells ◽  
Stem Cell ◽  
Cancer Stem Cells ◽  
Head And Neck ◽  
Cell Line ◽  
Cell Lines ◽  
Cancer Cell ◽  
Cancer Cell Lines

Purpose: Recent studies have shown that cancer stem cells are resistant to chemotherapy. The aim of this study was to compare RIF1 gene expression in head and neck, pancreatic cancer and glioma cell lines and the cancer stem cells isolated from these cell lines. Methods: UT-SCC-74 from Turku University and UT-SCC-74B primary tumor metastasis and neck cancer cell lines, YKG-1 glioma cancer cell line from RIKEN, pancreatic cancer cell lines and ASPC-1 cells from ATCC were grown in cell culture. To isolate cancer stem cells, ALDH-1 for UT-SCC-74 and UT-SCC-74B cell line, CD-133 for YKG-1 cell line and CD-24 for ASPC-1 cell line, were used as markers of cancer stem cells. RNA isolation was performed for both cancer lines and cancer stem cells. RNAs were converted to cDNA. RIF1 gene expression was performed by qRT-PCR analysis. RIF1 gene expression was compared with cancer cell lines and cancer stem cells isolated from these cell lines. The possible effect of RIF1 gene was evaluated. Results: In the pancreatic cells, RIF1 gene expression in the stem cell-positive cell line was 256 time that seen in the stem cell-negative cell line. Conclusion: Considering the importance of RIF1 in NHEJ and of NHEJ in pancreatic cancer, RIF1 may be one of the genes that plays an important role in the diagnoses and therapeutic treatment of pancreatic cancer. The results of head and neck and brain cancers are inconclusive and further studies are required to elucidate the connection between RIF1 gene and these other types of cancers.

scholarly journals A novel heterogeneous network-based method for drug response prediction in cancer cell lines

2018 ◽  
Vol 8 (1) ◽  
Author(s):  
Fei Zhang ◽  
Minghui Wang ◽  
Jianing Xi ◽  
Jianghong Yang ◽  
Ao Li
Keyword(s):  
Cell Lines ◽  
Cancer Cell ◽  
Heterogeneous Network ◽  
Drug Response ◽  
Response Prediction ◽  
Cancer Cell Lines

Metabolic evaluation of sporadic papillary kidney cancer.

2012 ◽  
Vol 30 (5_suppl) ◽  
pp. 377-377
Author(s):  
Brian Shuch ◽  
Christopher Ricketts ◽  
Carole Sourbier ◽  
Shinji Tsutsumi ◽  
Xiu-ying Zhang ◽  
...  
Keyword(s):  
Gene Expression ◽  
Membrane Potential ◽  
Cell Line ◽  
Cell Lines ◽  
Cancer Cell ◽  
Kidney Cancer ◽  
Expression Profiles ◽  
Fumarate Hydratase ◽  
Gene Expression Profiles ◽  
Cancer Cell Lines

377 Background: Papillary kidney cancer, which occurs in 15% of patients with kidney cancer, can be aggressive and there is currently no effective form of therapy for this disease. To evaluate the metabolic characteristics of sporadic papillary kidney cancer, we have evaluated metabolic parameters of several papillary kidney cancer cell lines and available gene expression profiles. Methods: Established cell lines derived from patients with sporadic papillary kidney cancer (LABAZ, MDACC-55, HRC-86T2) and from a hereditary form of fumarate hydratase-deficient kidney cancer (UOK262) were evaluated. All sporadic lines were initially sequenced for fumarate hydratase (FH). All cell lines were metabolically profiled using the Seahorse Extracellular Flux Analyzer and further evaluated for reactive oxygen species (ROS), mitochondrial membrane potential, and glucose dependence. Finally gene expression profiles of publically available datasets of papillary and HLRCC tumors were downloaded, normalized, and analyzed. Results: Sporadic lines had no alterations in FH and metabolic analysis demonstrated normal oxygen consumption and minimal lactate production, in contrast to highly glycolytic UOK262. Also unlike UOK262, the sporadic papillary kidney cancer lines were not sensitive to glucose withdrawal, had low levels of ROS, and had normal mitochondria membrane potential. Principal component analysis (PCA) demonstrated that HLRCC tumor specimens are very different from sporadic papillary tumors at the molecular level. Conclusions: Our study of established sporadic papillary RCC and fumarate hydratase-deficient HLRCC cell line together with analysis of available gene expression profiles demonstrates that these sporadic papillary kidney cancer cell lines appear to have a distinct metabolic profile from those in the fumarate hydratase deficient kidney cancer cell line. Understanding the metabolic basis of papillary kidney cancer could provide the foundation for the development of targeted approaches to therapy for patients with this disease.

Modular within and between score for drug response prediction in cancer cell lines

2020 ◽  
Vol 16 (1) ◽  
pp. 31-38
Author(s):  
Shiming Wang ◽  
Jie Li
Keyword(s):  
Cell Lines ◽  
Anticancer Drugs ◽  
Cancer Cell ◽  
Drug Response ◽  
Response Prediction ◽  
Cancer Cell Lines

Drug response prediction in cancer cell lines is vital to discover anticancer drugs for new cell lines.

scholarly journals RNA Sequencing of Hepatobiliary Cancer Cell Lines: Data and Applications to Mutational and Transcriptomic Profiling

Cancers ◽  
2020 ◽  
Vol 12 (9) ◽  
pp. 2510
Author(s):  
Dominique Scherer ◽  
Marcela Dávila López ◽  
Benjamin Goeppert ◽  
Sanna Abrahamsson ◽  
Rosa González Silos ◽  
...  
Keyword(s):  
Gene Expression ◽  
Cell Line ◽  
Rna Sequencing ◽  
Cell Lines ◽  
Cancer Cell ◽  
Rna Extraction ◽  
Cancer Cell Lines ◽  
Sequencing Data ◽  
Hepatobiliary Cancer ◽  
Public Data

Cancer cell lines allow the identification of clinically relevant alterations and the prediction of drug response. However, sequencing data for hepatobiliary cancer cell lines in general, and particularly gallbladder cancer (GBC), are sparse. Here, we apply RNA sequencing to characterize 10 GBC, eight hepatocellular carcinoma, and five cholangiocarcinoma (CCA) cell lines. RNA extraction, quality control, library preparation, sequencing, and pre-processing of sequencing data were implemented using state-of-the-art techniques. Public data from the MSK-IMPACT database and a large cohort of Japanese biliary tract cancer patients were used to illustrate the usage of the released data. The total number of exonic mutations varied from 7207 for the cell line NOZ to 9760 for HuCCT1. Researchers planning experiments that require TP53 mutations could use the cell lines NOZ, OCUG-1, SNU308, or YoMi. Mz-Cha-1 showed mutations in ATM, SNU308 presented SMAD4 mutations, and the only investigated cell line that showed ARID1A mutations was GB-d1. SNU478 was the cell line with the global gene expression pattern most similar to GBC, intrahepatic CCA, and extrahepatic CCA. EGFR, KMT2D, and KMT2C generally presented a higher expression in the investigated cell lines than in Japanese primary GBC tumors. We provide the scientific community with detailed mutation and gene expression data, together with three showcase applications, with the aim of facilitating the design of future in vitro cell culture assays for research on hepatobiliary cancer.

Sign in / Sign up

Export Citation Format

Share Document