Limits...
Drug target prioritization by perturbed gene expression and network information.

Isik Z, Baldow C, Cannistraci CV, Schroeder M - Sci Rep (2015)

Bottom Line: We performed the first systematic analysis of over 500 drugs from the Connectivity Map.We introduce a novel measure, local radiality, which combines perturbed genes and functional interaction network information.The new measure outperforms other methods in target prioritization and proposes cancer-specific pathways from drugs to affected genes for the first time.

View Article: PubMed Central - PubMed

Affiliation: Bioinformatics Group, Biotechnology Center (BIOTEC), Technische Universitat Dresden, Tatzberg 47-49, 01307 Dresden, Germany.

ABSTRACT
Drugs bind to their target proteins, which interact with downstream effectors and ultimately perturb the transcriptome of a cancer cell. These perturbations reveal information about their source, i.e., drugs' targets. Here, we investigate whether these perturbations and protein interaction networks can uncover drug targets and key pathways. We performed the first systematic analysis of over 500 drugs from the Connectivity Map. First, we show that the gene expression of drug targets is usually not significantly affected by the drug perturbation. Hence, expression changes after drug treatment on their own are not sufficient to identify drug targets. However, ranking of candidate drug targets by network topological measures prioritizes the targets. We introduce a novel measure, local radiality, which combines perturbed genes and functional interaction network information. The new measure outperforms other methods in target prioritization and proposes cancer-specific pathways from drugs to affected genes for the first time. Local radiality identifies more diverse targets with fewer neighbors and possibly less side effects.

No MeSH data available.


Related in: MedlinePlus

Results for drug-target prioritization methods.(A) Prediction performance of the selected measures for functional targets (FT). The y-axis shows the cumulative percentage of correctly predicted targets (i.e., recall) of all drugs in the CMap, the x-axis gives the predicted rank level. The predictions are given for the 1st percentile (top 120) of the ranking list. The LR achieved 22% recall value, which is the highest prediction rate. (B) The prediction power (expressed in decibel, dB) of each measure compared to the random predictor. It shows the magnitude of recall for each predictor normalized with respect to the random predictor. (C) The overlap of known targets that are predicted in the 1st percentile. 79 targets (common predicted) are predicted by all measures. Radiality and stress usually predict similar targets, however LR (136 unique targets) and kernel diffusion (77 unique targets) predict different ones. (D) The overlap of the drugs that bind to proteins found by only LR (LR Only) and all measures (Common Predicted). There were 331 different drugs that bind to 79 proteins, which are predicted by several measures. However, 15 drugs bind to specific proteins that are predicted only by LR. In other words, common targets are usually well-studied proteins, while the LR targets are more specific ones and have more potential for new drugs.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4663505&req=5

f3: Results for drug-target prioritization methods.(A) Prediction performance of the selected measures for functional targets (FT). The y-axis shows the cumulative percentage of correctly predicted targets (i.e., recall) of all drugs in the CMap, the x-axis gives the predicted rank level. The predictions are given for the 1st percentile (top 120) of the ranking list. The LR achieved 22% recall value, which is the highest prediction rate. (B) The prediction power (expressed in decibel, dB) of each measure compared to the random predictor. It shows the magnitude of recall for each predictor normalized with respect to the random predictor. (C) The overlap of known targets that are predicted in the 1st percentile. 79 targets (common predicted) are predicted by all measures. Radiality and stress usually predict similar targets, however LR (136 unique targets) and kernel diffusion (77 unique targets) predict different ones. (D) The overlap of the drugs that bind to proteins found by only LR (LR Only) and all measures (Common Predicted). There were 331 different drugs that bind to 79 proteins, which are predicted by several measures. However, 15 drugs bind to specific proteins that are predicted only by LR. In other words, common targets are usually well-studied proteins, while the LR targets are more specific ones and have more potential for new drugs.

Mentions: LR is the only measure combining expression and local network data based on shortest paths. We systematically compared all 13 measures (see Supplementary Fig. 2 for all measures). The performance of best predictors is shown in Fig. 3. Predictions based only on FC values are also included for better comparison. The LR performs best overall under all configurations. The random predictor ranges from 0.1 to 1% (see Fig. 3a). The LR predicts up to 22% of targets in its 1st percentile of predictions. Fig. 3b shows the performance of selected measures relative to a random predictor (prediction power). The symmetric kernel diffusion ranking is a successful metric developed recently4. Its recall value (0.17) is significantly (Wilcoxon signed rank test, p-value < 1.7e−16, Supplementary Table 3) lower than LR (0.22). The result is similar for the AUC value; the symmetric kernel diffusion ranking has an overall 0.81 AUC, and LR has 0.85 AUC (Supplementary Fig. 3a). The PeC was developed as the essential protein discovery measure8, and LR has a significantly (p-value < 5.7e−24) higher recall value versus PeC (0.17). PeC has a lower AUC (0.82) compared to LR (0.85). A recent study applied degree and betweenness to identify cancer targets9. In terms of recall values, LR significantly outperforms both degree (p-value < 1.2e−44) and betweenness (p-value < 2.7e−16). Similarly, LR has a higher AUC value compared to degree (0.81 AUC) and betweenness (0.80 AUC). In summary, LR performs significantly better (average p-value < 5.3e−06, Supplementary Table 3) than other predictors in the 1st percentile. The overall performance of LR also outperforms others with an AUC value of 0.85 (Supplementary Fig. 3a).


Drug target prioritization by perturbed gene expression and network information.

Isik Z, Baldow C, Cannistraci CV, Schroeder M - Sci Rep (2015)

Results for drug-target prioritization methods.(A) Prediction performance of the selected measures for functional targets (FT). The y-axis shows the cumulative percentage of correctly predicted targets (i.e., recall) of all drugs in the CMap, the x-axis gives the predicted rank level. The predictions are given for the 1st percentile (top 120) of the ranking list. The LR achieved 22% recall value, which is the highest prediction rate. (B) The prediction power (expressed in decibel, dB) of each measure compared to the random predictor. It shows the magnitude of recall for each predictor normalized with respect to the random predictor. (C) The overlap of known targets that are predicted in the 1st percentile. 79 targets (common predicted) are predicted by all measures. Radiality and stress usually predict similar targets, however LR (136 unique targets) and kernel diffusion (77 unique targets) predict different ones. (D) The overlap of the drugs that bind to proteins found by only LR (LR Only) and all measures (Common Predicted). There were 331 different drugs that bind to 79 proteins, which are predicted by several measures. However, 15 drugs bind to specific proteins that are predicted only by LR. In other words, common targets are usually well-studied proteins, while the LR targets are more specific ones and have more potential for new drugs.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4663505&req=5

f3: Results for drug-target prioritization methods.(A) Prediction performance of the selected measures for functional targets (FT). The y-axis shows the cumulative percentage of correctly predicted targets (i.e., recall) of all drugs in the CMap, the x-axis gives the predicted rank level. The predictions are given for the 1st percentile (top 120) of the ranking list. The LR achieved 22% recall value, which is the highest prediction rate. (B) The prediction power (expressed in decibel, dB) of each measure compared to the random predictor. It shows the magnitude of recall for each predictor normalized with respect to the random predictor. (C) The overlap of known targets that are predicted in the 1st percentile. 79 targets (common predicted) are predicted by all measures. Radiality and stress usually predict similar targets, however LR (136 unique targets) and kernel diffusion (77 unique targets) predict different ones. (D) The overlap of the drugs that bind to proteins found by only LR (LR Only) and all measures (Common Predicted). There were 331 different drugs that bind to 79 proteins, which are predicted by several measures. However, 15 drugs bind to specific proteins that are predicted only by LR. In other words, common targets are usually well-studied proteins, while the LR targets are more specific ones and have more potential for new drugs.
Mentions: LR is the only measure combining expression and local network data based on shortest paths. We systematically compared all 13 measures (see Supplementary Fig. 2 for all measures). The performance of best predictors is shown in Fig. 3. Predictions based only on FC values are also included for better comparison. The LR performs best overall under all configurations. The random predictor ranges from 0.1 to 1% (see Fig. 3a). The LR predicts up to 22% of targets in its 1st percentile of predictions. Fig. 3b shows the performance of selected measures relative to a random predictor (prediction power). The symmetric kernel diffusion ranking is a successful metric developed recently4. Its recall value (0.17) is significantly (Wilcoxon signed rank test, p-value < 1.7e−16, Supplementary Table 3) lower than LR (0.22). The result is similar for the AUC value; the symmetric kernel diffusion ranking has an overall 0.81 AUC, and LR has 0.85 AUC (Supplementary Fig. 3a). The PeC was developed as the essential protein discovery measure8, and LR has a significantly (p-value < 5.7e−24) higher recall value versus PeC (0.17). PeC has a lower AUC (0.82) compared to LR (0.85). A recent study applied degree and betweenness to identify cancer targets9. In terms of recall values, LR significantly outperforms both degree (p-value < 1.2e−44) and betweenness (p-value < 2.7e−16). Similarly, LR has a higher AUC value compared to degree (0.81 AUC) and betweenness (0.80 AUC). In summary, LR performs significantly better (average p-value < 5.3e−06, Supplementary Table 3) than other predictors in the 1st percentile. The overall performance of LR also outperforms others with an AUC value of 0.85 (Supplementary Fig. 3a).

Bottom Line: We performed the first systematic analysis of over 500 drugs from the Connectivity Map.We introduce a novel measure, local radiality, which combines perturbed genes and functional interaction network information.The new measure outperforms other methods in target prioritization and proposes cancer-specific pathways from drugs to affected genes for the first time.

View Article: PubMed Central - PubMed

Affiliation: Bioinformatics Group, Biotechnology Center (BIOTEC), Technische Universitat Dresden, Tatzberg 47-49, 01307 Dresden, Germany.

ABSTRACT
Drugs bind to their target proteins, which interact with downstream effectors and ultimately perturb the transcriptome of a cancer cell. These perturbations reveal information about their source, i.e., drugs' targets. Here, we investigate whether these perturbations and protein interaction networks can uncover drug targets and key pathways. We performed the first systematic analysis of over 500 drugs from the Connectivity Map. First, we show that the gene expression of drug targets is usually not significantly affected by the drug perturbation. Hence, expression changes after drug treatment on their own are not sufficient to identify drug targets. However, ranking of candidate drug targets by network topological measures prioritizes the targets. We introduce a novel measure, local radiality, which combines perturbed genes and functional interaction network information. The new measure outperforms other methods in target prioritization and proposes cancer-specific pathways from drugs to affected genes for the first time. Local radiality identifies more diverse targets with fewer neighbors and possibly less side effects.

No MeSH data available.


Related in: MedlinePlus