Structural Chemogenomics

Briels, Babs; Graaf, Chris de; Bender, Andreas

doi:10.1002/9781118681121.ch3

Cited by 3 publications

(4 citation statements)

References 133 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While the sparse set is only 15% bigger than the dense set, its small molecules number is ~74 times more prominent ( Fig 3B ). The densely-covered test set represents a typical chemogenomics [ 1 , 82 , 83 ] or proteochemometrics [ 8 , 9 ] matrix where each small molecule is experimentally validated on at least 60 kinases (out of 69 in total). The structures of the 278 inhibitors of the dense set can be found in S3 Table .…”

Section: Resultsmentioning

confidence: 99%

“…The average cost of developing a new drug has increased by a factor of 9 in the last 50 years. [1] And although personalized therapies are steadily showing their value in clinical trials, [2,3] these trials can suffer from therapy resistance, [4][5][6] requiring adaptive treatment strategies. Network-based and proteochemometrics approaches [7][8][9] introduced a decade ago offer a paradigm shift from highly selective single-target drugs to multi-target drugs where the latter is more likely to achieve desired clinical efficacy and potentially lower the research and development (R&D) costs.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Predicting the target landscape of kinase inhibitors using 3D convolutional neural networks

Kanev,

Zhang,

Kooistra

et al. 2023

PLoS Comput Biol

Self Cite

View full text Add to dashboard Cite

Many therapies in clinical trials are based on single drug-single target relationships. To further extend this concept to multi-target approaches using multi-targeted drugs, we developed a machine learning pipeline to unravel the target landscape of kinase inhibitors. This pipeline, which we call 3D-KINEssence, uses a new type of protein fingerprints (3D FP) based on the structure of kinases generated through a 3D convolutional neural network (3D-CNN). These 3D-CNN kinase fingerprints were matched to molecular Morgan fingerprints to predict the targets of each respective kinase inhibitor based on available bioactivity data. The performance of the pipeline was evaluated on two test sets: a sparse drug-target set where each drug is matched in most cases to a single target and also on a densely-covered drug-target set where each drug is matched to most if not all targets. This latter set is more challenging to train, given its non-exclusive character. Our model’s root-mean-square error (RMSE) based on the two datasets was 0.68 and 0.8, respectively. These results indicate that 3D FP can predict the target landscape of kinase inhibitors at around 0.8 log units of bioactivity. Our strategy can be utilized in proteochemometric or chemogenomic workflows by consolidating the target landscape of kinase inhibitors.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Predicting the target landscape of kinase inhibitors using 3D convolutional neural networks

Kanev,

Zhang,

Kooistra

et al. 2023

PLoS Comput Biol

Self Cite

View full text Add to dashboard Cite

show abstract

“…The average cost of developing a new drug has increased by a factor of 9 in the last 50 years. [402] And although personalized therapies are steadily showing their value in clinical trials, [30,31] these trials can suffer from therapy resistance, [3,32,33] requiring adaptive treatment strategies. Network-based and proteochemometrics approaches [82,176,403] introduced a decade ago offer a paradigm shift from highly selective single-target drugs to multi-target drugs where the latter is more likely to achieve desired clinical efficacy and potentially lower the research and development (R&D) costs.…”

Section: Chaptermentioning

confidence: 99%

“…While the sparse set is only 15% bigger than the dense set, its small molecules number is ~74 times more prominent (Figure 5.3). The densely-covered test set represents a typical chemogenomics [145,153,402] or proteochemometrics [82,176] matrix…”

Section: Protein Features Contribute To Better Scoring Of Machine Lea...mentioning

confidence: 99%

Integration of Phenotypic Drug Efficacy and Molecular Chemogenomics Data

Kanev

View full text Add to dashboard Cite

This thesis thoroughly analyzed the sequence, structure, and molecular interactions of the two major classes of protein kinases - eukaryotic and atypical. This analysis was intended to enhance our understanding of inhibitor binding and selectivity and to create advanced machine-learning pipelines capable of leveraging these differences to predict small molecules with desired polypharmacological characteristics.To reach this aim, the following objectives were defined: 1). Extend the KLIFS database with the atypical kinase class. 2). Systematically analyze the sequence, structural, and molecular interaction kinase data from KLIFS for new insights into inhibitor binding and specificity. 3). Integrate and curate bioactivity data from ChEMBL, KIEO, literature, and genomics data from COSMIC, dbSNP, and cBioPortal. 4). Develop a convolutional neural network pipeline that can learn from 3D kinase structures to predict the polypharmacological profiles of small- molecule protein kinase inhibitors.

show abstract