Combined kinome inhibition states are predictive of cancer cell line sensitivity to kinase inhibitor combination therapies DOI Creative Commons
Chinmaya U. Joisa, Kevin A. Chen,

Samantha Beville

et al.

Biocomputing, Journal Year: 2023, Volume and Issue: unknown, P. 276 - 290

Published: Dec. 1, 2023

Protein kinases are a primary focus in targeted therapy development for cancer, owing to their role as regulators nearly all areas of cell life. Recent strategies targeting the kinome with combination therapies have shown promise, such trametinib and dabrafenib advanced melanoma, but empirical design less characterized pathways remains challenge. Computational screening is an attractive alternative, allowing in-silico filtering prior experimental testing drastically fewer leads, increasing efficiency effectiveness drug pipelines. In this work, we generated combined inhibition states 40,000 kinase inhibitor combinations from kinobeads-based profiling across 64 doses. We then integrated these transcriptomics CCLE build machine learning models elastic-net feature selection predict line sensitivity nine cancer types, accuracy R2 ∼ 0.75-0.9. validated model by using PDX-derived TNBC saw good global (R2 0.7) well high predicting synergy four popular metrics 0.9). Additionally, was able highly synergistic omipalisib treatment, which incidentally recently phase I clinical trials. Our choice tree-based greater interpretability allowed interrogation predictive each type, MAPK, CDK, STK kinases. Overall, results suggest that strongly responses great potential integration into computational This approach may facilitate identification effective accelerate novel therapies, ultimately improving patient outcomes.

Language: Английский

Integrated single-dose kinome profiling data is predictive of cancer cell line sensitivity to kinase inhibitors DOI Creative Commons
Chinmaya U. Joisa, Kevin A. Chen, Matthew E. Berginski

et al.

PeerJ, Journal Year: 2023, Volume and Issue: 11, P. e16342 - e16342

Published: Nov. 16, 2023

Protein kinase activity forms the backbone of cellular information transfer, acting both individually and as part a broader network, kinome. Their central role in signaling leads to kinome dysfunction being common driver disease, particular cancer, where numerous kinases have been identified having causal or modulating tumor development progression. As result, therapies targeting has rapidly grown, with over 70 inhibitors approved for use clinic double this number currently clinical trials. Understanding relationship between inhibitor treatment their effects on downstream phenotype is thus clear importance understanding mechanisms streamlining compound screening therapy development. In work, we combine two large-scale profiling data sets them link inhibitor-kinome interactions cell line responses (AUC/IC 50 ). We then built computational models set that achieve high degree prediction accuracy (R 2 0.7 RMSE 0.9) were able identify well-characterized understudied significantly affect responses. further validated these experimentally by testing predicted breast cancer lines extended model scope performing additional validation patient-derived pancreatic lines. Overall, results demonstrate broad quantification inhibition state highly predictive phenotypes.

Language: Английский

Citations

4
Combined kinome inhibition states are predictive of cancer cell line sensitivity to kinase inhibitor combination therapies DOI Creative Commons
Chinmaya U. Joisa, Kevin A. Chen,

Samantha Beville

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: Aug. 3, 2023

Protein kinases are a primary focus in targeted therapy development for cancer, owing to their role as regulators nearly all areas of cell life. Kinase inhibitors one the fastest growing drug classes oncology, but resistance acquisition kinase-targeting monotherapies is inevitable due dynamic and interconnected nature kinome response perturbation. Recent strategies targeting with combination therapies have shown promise, such approval Trametinib Dabrafenib advanced melanoma, similar empirical design less characterized pathways remains challenge. Computational screening an attractive alternative, allowing in-silico prior in-vitro or in-vivo testing drastically fewer leads, increasing efficiency effectiveness pipelines. In this work, we generate combined inhibition states 40,000 kinase inhibitor combinations from kinobeads-based profiling across 64 doses. We then integrated these baseline transcriptomics CCLE build robust machine learning models predict line sensitivity NCI-ALMANAC nine cancer types, model accuracy R

Language: Английский

Citations

0
Combined kinome inhibition states are predictive of cancer cell line sensitivity to kinase inhibitor combination therapies DOI Creative Commons
Chinmaya U. Joisa, Kevin A. Chen,

Samantha Beville

et al.

Biocomputing, Journal Year: 2023, Volume and Issue: unknown, P. 276 - 290

Published: Dec. 1, 2023

Protein kinases are a primary focus in targeted therapy development for cancer, owing to their role as regulators nearly all areas of cell life. Recent strategies targeting the kinome with combination therapies have shown promise, such trametinib and dabrafenib advanced melanoma, but empirical design less characterized pathways remains challenge. Computational screening is an attractive alternative, allowing in-silico filtering prior experimental testing drastically fewer leads, increasing efficiency effectiveness drug pipelines. In this work, we generated combined inhibition states 40,000 kinase inhibitor combinations from kinobeads-based profiling across 64 doses. We then integrated these transcriptomics CCLE build machine learning models elastic-net feature selection predict line sensitivity nine cancer types, accuracy R2 ∼ 0.75-0.9. validated model by using PDX-derived TNBC saw good global (R2 0.7) well high predicting synergy four popular metrics 0.9). Additionally, was able highly synergistic omipalisib treatment, which incidentally recently phase I clinical trials. Our choice tree-based greater interpretability allowed interrogation predictive each type, MAPK, CDK, STK kinases. Overall, results suggest that strongly responses great potential integration into computational This approach may facilitate identification effective accelerate novel therapies, ultimately improving patient outcomes.

Language: Английский

Citations

0