Limits...
Predicting Anticancer Drug Responses Using a Dual-Layer Integrated Cell Line-Drug Network Model.

Zhang N, Wang H, Fang Y, Wang J, Zheng X, Liu XS - PLoS Comput. Biol. (2015)

Bottom Line: When using the dual-layer model integrating both CSN and DSN, our predicted response reached a 0.6 Pearson correlation coefficient with observed responses for most drugs, which is significantly better than the previous results using the elastic net model.We have also applied the dual-layer cell line-drug integrated network model to fill in the missing drug response values in the CGP dataset.Even though the dual-layer integrated cell line-drug network model does not specifically model mutation information, it correctly predicted that BRAF mutant cell lines would be more sensitive than BRAF wild-type cell lines to three MEK1/2 inhibitors tested.

View Article: PubMed Central - PubMed

Affiliation: Department of Mathematics, Shanghai Normal University, Shanghai, China.

ABSTRACT
The ability to predict the response of a cancer patient to a therapeutic agent is a major goal in modern oncology that should ultimately lead to personalized treatment. Existing approaches to predicting drug sensitivity rely primarily on profiling of cancer cell line panels that have been treated with different drugs and selecting genomic or functional genomic features to regress or classify the drug response. Here, we propose a dual-layer integrated cell line-drug network model, which uses both cell line similarity network (CSN) data and drug similarity network (DSN) data to predict the drug response of a given cell line using a weighted model. Using the Cancer Cell Line Encyclopedia (CCLE) and Cancer Genome Project (CGP) studies as benchmark datasets, our single-layer model with CSN or DSN and only a single parameter achieved a prediction performance comparable to the previously generated elastic net model. When using the dual-layer model integrating both CSN and DSN, our predicted response reached a 0.6 Pearson correlation coefficient with observed responses for most drugs, which is significantly better than the previous results using the elastic net model. We have also applied the dual-layer cell line-drug integrated network model to fill in the missing drug response values in the CGP dataset. Even though the dual-layer integrated cell line-drug network model does not specifically model mutation information, it correctly predicted that BRAF mutant cell lines would be more sensitive than BRAF wild-type cell lines to three MEK1/2 inhibitors tested.

No MeSH data available.


Related in: MedlinePlus

Computation framework.(A–E) Construction of the dual-layer integrated cell line-drug network model. (F–H) Drug sensitivity predictions based on information from the integrated cell line-drug network. (F) Two sub-networks for predicting the response of cell line C to drug D. (G) For each sub-network, the weighted arithmetic mean was used to measure an unknown drug-cell line response based on their neighboring cell lines or drugs. (H) Predictions from each of the two individual sub-networks were merged using a weighted model to determine the final prediction.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4587957&req=5

pcbi.1004498.g002: Computation framework.(A–E) Construction of the dual-layer integrated cell line-drug network model. (F–H) Drug sensitivity predictions based on information from the integrated cell line-drug network. (F) Two sub-networks for predicting the response of cell line C to drug D. (G) For each sub-network, the weighted arithmetic mean was used to measure an unknown drug-cell line response based on their neighboring cell lines or drugs. (H) Predictions from each of the two individual sub-networks were merged using a weighted model to determine the final prediction.

Mentions: Based on the above results, we developed a dual-layer integrated cell line-drug network model to predict anticancer drug sensitivity using existing cancer cell line expression profiles and drug response data (Fig 2). The model integrated three types of data: 1) gene expression profiles for each cell line; 2) 1-D and 2-D chemical structural properties of each drug; and 3) the drug response for each cell line. The top layer of the network, termed cell line similarity network (CSN), predicts the response of cell line C to a given drug D using a linear model weighting drug response from cell lines with similar gene expression profiles to C. We calculated the gene expression correlations between C with all other cell lines (Fig 2A and 2B), and gave higher weights to more similar ones. The bottom layer, termed drug similarity network (DSN), predicts the response of cell line C to drug D weighting the response data on drugs similar to D in their chemical structures. We calculated the correlation of every drug pair based on their 1-D and 2-D chemical structure features in PubChem. The two layers, CSN and DSN, were connected using drug response data for the cell lines, which were represented as activity areas in CCLE [13] and CGP studies [14]. Notably, the network is not a complete bipartite graph, as some drug response data for some cell lines are missing in these studies, especially for the CGP dataset.


Predicting Anticancer Drug Responses Using a Dual-Layer Integrated Cell Line-Drug Network Model.

Zhang N, Wang H, Fang Y, Wang J, Zheng X, Liu XS - PLoS Comput. Biol. (2015)

Computation framework.(A–E) Construction of the dual-layer integrated cell line-drug network model. (F–H) Drug sensitivity predictions based on information from the integrated cell line-drug network. (F) Two sub-networks for predicting the response of cell line C to drug D. (G) For each sub-network, the weighted arithmetic mean was used to measure an unknown drug-cell line response based on their neighboring cell lines or drugs. (H) Predictions from each of the two individual sub-networks were merged using a weighted model to determine the final prediction.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4587957&req=5

pcbi.1004498.g002: Computation framework.(A–E) Construction of the dual-layer integrated cell line-drug network model. (F–H) Drug sensitivity predictions based on information from the integrated cell line-drug network. (F) Two sub-networks for predicting the response of cell line C to drug D. (G) For each sub-network, the weighted arithmetic mean was used to measure an unknown drug-cell line response based on their neighboring cell lines or drugs. (H) Predictions from each of the two individual sub-networks were merged using a weighted model to determine the final prediction.
Mentions: Based on the above results, we developed a dual-layer integrated cell line-drug network model to predict anticancer drug sensitivity using existing cancer cell line expression profiles and drug response data (Fig 2). The model integrated three types of data: 1) gene expression profiles for each cell line; 2) 1-D and 2-D chemical structural properties of each drug; and 3) the drug response for each cell line. The top layer of the network, termed cell line similarity network (CSN), predicts the response of cell line C to a given drug D using a linear model weighting drug response from cell lines with similar gene expression profiles to C. We calculated the gene expression correlations between C with all other cell lines (Fig 2A and 2B), and gave higher weights to more similar ones. The bottom layer, termed drug similarity network (DSN), predicts the response of cell line C to drug D weighting the response data on drugs similar to D in their chemical structures. We calculated the correlation of every drug pair based on their 1-D and 2-D chemical structure features in PubChem. The two layers, CSN and DSN, were connected using drug response data for the cell lines, which were represented as activity areas in CCLE [13] and CGP studies [14]. Notably, the network is not a complete bipartite graph, as some drug response data for some cell lines are missing in these studies, especially for the CGP dataset.

Bottom Line: When using the dual-layer model integrating both CSN and DSN, our predicted response reached a 0.6 Pearson correlation coefficient with observed responses for most drugs, which is significantly better than the previous results using the elastic net model.We have also applied the dual-layer cell line-drug integrated network model to fill in the missing drug response values in the CGP dataset.Even though the dual-layer integrated cell line-drug network model does not specifically model mutation information, it correctly predicted that BRAF mutant cell lines would be more sensitive than BRAF wild-type cell lines to three MEK1/2 inhibitors tested.

View Article: PubMed Central - PubMed

Affiliation: Department of Mathematics, Shanghai Normal University, Shanghai, China.

ABSTRACT
The ability to predict the response of a cancer patient to a therapeutic agent is a major goal in modern oncology that should ultimately lead to personalized treatment. Existing approaches to predicting drug sensitivity rely primarily on profiling of cancer cell line panels that have been treated with different drugs and selecting genomic or functional genomic features to regress or classify the drug response. Here, we propose a dual-layer integrated cell line-drug network model, which uses both cell line similarity network (CSN) data and drug similarity network (DSN) data to predict the drug response of a given cell line using a weighted model. Using the Cancer Cell Line Encyclopedia (CCLE) and Cancer Genome Project (CGP) studies as benchmark datasets, our single-layer model with CSN or DSN and only a single parameter achieved a prediction performance comparable to the previously generated elastic net model. When using the dual-layer model integrating both CSN and DSN, our predicted response reached a 0.6 Pearson correlation coefficient with observed responses for most drugs, which is significantly better than the previous results using the elastic net model. We have also applied the dual-layer cell line-drug integrated network model to fill in the missing drug response values in the CGP dataset. Even though the dual-layer integrated cell line-drug network model does not specifically model mutation information, it correctly predicted that BRAF mutant cell lines would be more sensitive than BRAF wild-type cell lines to three MEK1/2 inhibitors tested.

No MeSH data available.


Related in: MedlinePlus