Systematic identification of genomic markers of drug sensitivity in cancer cells

Garnett, Mathew J.; Edelman, Elena J.; Heidorn, Sonja J.; Greenman, Christopher; Dastur, Anahita; Lau, King Wai; Greninger, Patricia; Thompson, Iain; Luo, Xi; Soares, Jorge; Liu, Qingsong; Iorio, Francesco; Surdez, Didier; Chen, Li; Milano, Randy; Bignell, Graham R.; Tam, Ah Ting; Davies, Helen; Stevenson, Jesse A.; Barthorpe, Syd; Lutz, Stephen; Kogera, Fiona; Lawrence, Karl P.; McLaren-Douglas, Anne; Mitropoulos, Xeni; Mironenko, Tatiana; Thi, Helen; Richardson, Laura L.; Zhou, Wenjun; Jewitt, F; Zhang, Tinghu; O’Brien, Patrick; Boisvert, Jessica L.; Price, Stacey; Hur, Wooyoung; Yang, Wanjuan; Deng, Xianming; Butler, Adam; Choi, Hwan Geun; Chang, Jae Won; Baselga, J.; Stamenkovic, Ivan; Engelman, Jeffrey A.; Sharma, Sreenath V.; Delattre, Olivier; Sáez-Rodríguez, Julio; Gray, Nathanael S.; Settleman, Jeffrey; Futreal, P. Andrew; Haber, Daniel A.; Stratton, Michael R.; Ramaswamy, Sridhar; McDermott, Ultan; Benes, Cyril H.

doi:10.1038/nature11005

Cited by 2,205 publications

(2,524 citation statements)

References 38 publications

Supporting

Mentioning

2,393

Contrasting

Unclassified

Order By: Relevance

“…Genomics of Drug Sensitivity in Cancer (GDSC) 8 , Cancer Cell Line Encyclopedia (CCLE) 9 , and profiling a panel of 60 cancer cell lines from the National Cancer Institute (NCI-60) 7 . Among them, GDSC data is currently offering the highest number of cell lines tested per drug 8 .…”

Section: Introductionmentioning

confidence: 99%

Systematic assessment of multi-gene predictors of pan-cancer cell line sensitivity to drugs exploiting gene expression data

Nguyen¹,

Dang²,

Ballester³

2016

F1000Res

View full text Add to dashboard Cite

Background: Selected gene mutations are routinely used to guide the selection of cancer drugs for a given patient tumour. Large pharmacogenomic data sets were introduced to discover more of these single-gene markers of drug sensitivity. Very recently, machine learning regression has been used to investigate how well cancer cell line sensitivity to drugs is predicted depending on the type of molecular profile. The latter has revealed that gene expression data is the most predictive profile in the pan-cancer setting. However, no study to date has exploited GDSC data to systematically compare the performance of machine learning models based on multi-gene expression data against that of widely-used single-gene markers based on genomics data. Methods: Here we present this systematic comparison using Random Forest (RF) classifiers exploiting the expression levels of 13,321 genes and an average of 501 tested cell lines per drug. To account for time-dependent batch effects in IC 50 measurements, we employ independent test sets generated with more recent GDSC data than that used to train the predictors and show that this is a more realistic validation than K-fold cross-validation. Results and Discussion: Across 127 GDSC drugs, our results show that the single-gene markers unveiled by the MANOVA analysis tend to achieve higher precision than these RF-based multi-gene models, at the cost of generally having a poor recall (i.e. correctly detecting only a small part of the cell lines sensitive to the drug). Regarding overall classification performance, about two thirds of the drugs are better predicted by multi-gene RF classifiers. Among the drugs with the most predictive of these models, we found pyrimethamine, sunitinib and 17-AAG. Conclusions: We now know that this type of models can predict in vitro tumour response to these drugs. These models can thus be further investigated on in vivo tumour models.

show abstract

Section: Introductionmentioning

confidence: 99%

Systematic assessment of multi-gene predictors of pan-cancer cell line sensitivity to drugs exploiting gene expression data

Nguyen¹,

Dang²,

Ballester³

2016

F1000Res

View full text Add to dashboard Cite

show abstract

“…Cell lines represent a mainstay to functionalize molecular data as they allow experimental manipulation, global and detailed mechanistic studies and highthroughput applications [1][2][3][4] . Virtually all commonly used cancer cells were continuously grown in vitro for years, and decades have often passed since they were originally derived from patients.…”

mentioning

confidence: 99%

The molecular landscape of colorectal cancer cell lines unveils clinically actionable kinase targets

et al. 2015

View full text Add to dashboard Cite

The development of molecularly targeted anticancer agents relies on large panels of tumourspecific preclinical models closely recapitulating the molecular heterogeneity observed in patients. Here we describe the mutational and gene expression analyses of 151 colorectal cancer (CRC) cell lines. We find that the whole spectrum of CRC molecular and transcriptional subtypes, previously defined in patients, is represented in this cell line compendium. Transcriptional outlier analysis identifies RAS/BRAF wild-type cells, resistant to EGFR blockade, functionally and pharmacologically addicted to kinase genes including ALK, FGFR2, NTRK1/2 and RET. The same genes are present as expression outliers in CRC patient samples. Genomic rearrangements (translocations) involving the ALK and NTRK1 genes are associated with the overexpression of the corresponding proteins in CRC specimens. The approach described here can be used to pinpoint CRCs with exquisite dependencies to individual kinases for which clinically approved drugs are already available.

show abstract

“…In order to identify individual genes whose expression profiles matched that of the one generated by cytotoxicity experiments for bortezomib, we used a linear regression-based approach, where we searched for statistically significant correlations between gene expression values and IC50 data [17,18,19]. The intersections of the genes were identified in 7 cell lines and used for further analysis.…”

Section: Resultsmentioning

confidence: 99%

“…We used an in silico literature mining approach to identify potential biomarkers by creating a summarized set of metadata derived from relevant information [17,18,19]. To do that, a linear regression model was used to discover genes whose expression profiles correlated with bortezomib sensitivity as measured for 7 myeloma cell lines by IC50 values from the Genomics of Drug Sensitivity in Cancer database.…”

Section: Methodsmentioning

confidence: 99%

Expression Profiles of the Individual Genes Corresponding to the Genes Generated by Cytotoxicity Experiments with Bortezomib in Multiple Myeloma

Ghasemi¹,

Alpsoy²,

Turk³

et al. 2016

Tjh

View full text Add to dashboard Cite

Objective:Multiple myeloma (MM) is currently incurable due to refractory disease relapse even under novel anti-myeloma treatment. In silico studies are effective for key decision making during clinicopathological battles against the chronic course of MM. The aim of this present in silico study was to identify individual genes whose expression profiles match that of the one generated by cytotoxicity experiments for bortezomib.Materials and Methods:We used an in silico literature mining approach to identify potential biomarkers by creating a summarized set of metadata derived from relevant information. The E-MTAB-783 dataset containing expression data from 789 cancer cell lines including 8 myeloma cell lines with drug screening data from the Wellcome Trust Sanger Institute database obtained from ArrayExpress was “Robust Multi-array analysis” normalized using GeneSpring v.12.5. Drug toxicity data were obtained from the Genomics of Drug Sensitivity in Cancer project. In order to identify individual genes whose expression profiles matched that of the one generated by cytotoxicity experiments for bortezomib, we used a linear regression-based approach, where we searched for statistically significant correlations between gene expression values and IC50 data. The intersections of the genes were identified in 8 cell lines and used for further analysis.Results:Our linear regression model identified 73 genes and some genes expression levels were found to very closely correlated with bortezomib IC50 values. When all 73 genes were used in a hierarchical cluster analysis, two major clusters of cells representing relatively sensitive and resistant cells could be identified. Pathway and molecular function analysis of all the significant genes was also investigated, as well as the genes involved in pathways.Conclusion:The findings of our present in silico study could be important not only for the understanding of the genomics of MM but also for the better arrangement of the targeted anti-myeloma therapies, such as bortezomib.

show abstract

Systematic identification of genomic markers of drug sensitivity in cancer cells

Cited by 2,205 publications

References 38 publications

Systematic assessment of multi-gene predictors of pan-cancer cell line sensitivity to drugs exploiting gene expression data

Systematic assessment of multi-gene predictors of pan-cancer cell line sensitivity to drugs exploiting gene expression data

The molecular landscape of colorectal cancer cell lines unveils clinically actionable kinase targets

Expression Profiles of the Individual Genes Corresponding to the Genes Generated by Cytotoxicity Experiments with Bortezomib in Multiple Myeloma

Contact Info

Product

Resources

About