Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity Prediction
Large-scale databases that report the inhibitory capacities of many combinations of candidate drug compounds and cultivated cancer cell lines have driven the development of preclinical drug-sensitivity models based on machine learning. However, cultivated cell lines have devolved from human cancer c...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-08-01
|
| Series: | Cancers |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-6694/14/16/3950 |
| _version_ | 1797446562351153152 |
|---|---|
| author | Paul Prasse Pascal Iversen Matthias Lienhard Kristina Thedinga Ralf Herwig Tobias Scheffer |
| author_facet | Paul Prasse Pascal Iversen Matthias Lienhard Kristina Thedinga Ralf Herwig Tobias Scheffer |
| author_sort | Paul Prasse |
| collection | DOAJ |
| description | Large-scale databases that report the inhibitory capacities of many combinations of candidate drug compounds and cultivated cancer cell lines have driven the development of preclinical drug-sensitivity models based on machine learning. However, cultivated cell lines have devolved from human cancer cells over years or even decades under selective pressure in culture conditions. Moreover, models that have been trained on in vitro data cannot account for interactions with other types of cells. Drug-response data that are based on patient-derived cell cultures, xenografts, and organoids, on the other hand, are not available in the quantities that are needed to train high-capacity machine-learning models. We found that pre-training deep neural network models of drug sensitivity on in vitro drug-sensitivity databases before fine-tuning the model parameters on patient-derived data improves the models’ accuracy and improves the biological plausibility of the features, compared to training only on patient-derived data. From our experiments, we can conclude that pre-trained models outperform models that have been trained on the target domains in the vast majority of cases. |
| first_indexed | 2024-03-09T13:43:18Z |
| format | Article |
| id | doaj.art-7689901b9a3940e7900c377f9965cc8d |
| institution | Directory Open Access Journal |
| issn | 2072-6694 |
| language | English |
| last_indexed | 2024-03-09T13:43:18Z |
| publishDate | 2022-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Cancers |
| spelling | doaj.art-7689901b9a3940e7900c377f9965cc8d2023-11-30T21:04:22ZengMDPI AGCancers2072-66942022-08-011416395010.3390/cancers14163950Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity PredictionPaul Prasse0Pascal Iversen1Matthias Lienhard2Kristina Thedinga3Ralf Herwig4Tobias Scheffer5Department of Computer Science, University of Potsdam, 14476 Potsdam, GermanyDepartment of Computer Science, University of Potsdam, 14476 Potsdam, GermanyDepartment of Computational Molecular Biology, Max Planck Institute for Molecular Genetics, 14195 Berlin, GermanyDepartment of Computational Molecular Biology, Max Planck Institute for Molecular Genetics, 14195 Berlin, GermanyDepartment of Computational Molecular Biology, Max Planck Institute for Molecular Genetics, 14195 Berlin, GermanyDepartment of Computer Science, University of Potsdam, 14476 Potsdam, GermanyLarge-scale databases that report the inhibitory capacities of many combinations of candidate drug compounds and cultivated cancer cell lines have driven the development of preclinical drug-sensitivity models based on machine learning. However, cultivated cell lines have devolved from human cancer cells over years or even decades under selective pressure in culture conditions. Moreover, models that have been trained on in vitro data cannot account for interactions with other types of cells. Drug-response data that are based on patient-derived cell cultures, xenografts, and organoids, on the other hand, are not available in the quantities that are needed to train high-capacity machine-learning models. We found that pre-training deep neural network models of drug sensitivity on in vitro drug-sensitivity databases before fine-tuning the model parameters on patient-derived data improves the models’ accuracy and improves the biological plausibility of the features, compared to training only on patient-derived data. From our experiments, we can conclude that pre-trained models outperform models that have been trained on the target domains in the vast majority of cases.https://www.mdpi.com/2072-6694/14/16/3950deep neural networksdrug-sensitivity predictionanti-cancer drugs |
| spellingShingle | Paul Prasse Pascal Iversen Matthias Lienhard Kristina Thedinga Ralf Herwig Tobias Scheffer Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity Prediction Cancers deep neural networks drug-sensitivity prediction anti-cancer drugs |
| title | Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity Prediction |
| title_full | Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity Prediction |
| title_fullStr | Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity Prediction |
| title_full_unstemmed | Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity Prediction |
| title_short | Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity Prediction |
| title_sort | pre training on in vitro and fine tuning on patient derived data improves deep neural networks for anti cancer drug sensitivity prediction |
| topic | deep neural networks drug-sensitivity prediction anti-cancer drugs |
| url | https://www.mdpi.com/2072-6694/14/16/3950 |
| work_keys_str_mv | AT paulprasse pretrainingoninvitroandfinetuningonpatientderiveddataimprovesdeepneuralnetworksforanticancerdrugsensitivityprediction AT pascaliversen pretrainingoninvitroandfinetuningonpatientderiveddataimprovesdeepneuralnetworksforanticancerdrugsensitivityprediction AT matthiaslienhard pretrainingoninvitroandfinetuningonpatientderiveddataimprovesdeepneuralnetworksforanticancerdrugsensitivityprediction AT kristinathedinga pretrainingoninvitroandfinetuningonpatientderiveddataimprovesdeepneuralnetworksforanticancerdrugsensitivityprediction AT ralfherwig pretrainingoninvitroandfinetuningonpatientderiveddataimprovesdeepneuralnetworksforanticancerdrugsensitivityprediction AT tobiasscheffer pretrainingoninvitroandfinetuningonpatientderiveddataimprovesdeepneuralnetworksforanticancerdrugsensitivityprediction |