Predicting patient response with models trained on cell lines and patient-derived xenografts by nonlinear transfer learning

被引:26
作者
Mourragui, Soufiane M. C. [1 ,2 ]
Loog, Marco [2 ,3 ]
Vis, Daniel J. [1 ]
Moore, Kat [1 ]
Manjon, Anna G. [4 ]
van de Wiel, Mark A. [5 ,6 ]
Reinders, Marcel J. T. [2 ,7 ]
Wessels, Lodewyk F. A. [1 ,2 ]
机构
[1] Netherlands Canc Inst, Oncode Inst, Div Mol Carcinogenesis, NL-1066 CX Amsterdam, Netherlands
[2] Delft Univ Technol, Dept Elect Engn Math & Comp Sci, NL-2628 XE Delft, Netherlands
[3] Univ Copenhagen, Dept Comp Sci, DK-2100 Copenhagen, Denmark
[4] Netherlands Canc Inst, Oncode Inst, Div Cell Biol, NL-1066 CX Amsterdam, Netherlands
[5] Amsterdam Univ Med Ctr, Epidemiol & Biostat, NL-1105 AZ Amsterdam, Netherlands
[6] Univ Cambridge, Biostat Unit, Med Res Council, Cambridge CB2 0SR, England
[7] Leiden Univ, Leiden Computat Biol Ctr, Med Ctr, NL-2333 ZC Leiden, Netherlands
关键词
model systems; translational medicine; clinical drug response; cancer; transfer learning; GENE-EXPRESSION; DRUG RESPONSE; CANCER; KERNEL; RESISTANCE; PACLITAXEL; INHIBITION; PROTEIN-1; RELEVANCE; CDC42;
D O I
10.1073/pnas.2106682118
中图分类号
O [数理科学和化学]; P [天文学、地球科学]; Q [生物科学]; N [自然科学总论];
学科分类号
07 ; 0710 ; 09 ;
摘要
Preclinical models have been the workhorse of cancer research, producing massive amounts of drug response data. Unfortunately, translating response biomarkers derived from these datasets to human tumors has proven to be particularly challenging. To address this challenge, we developed TRANSACT, a computational framework that builds a consensus space to capture biological processes common to preclinical models and human tumors and exploits this space to construct drug response predictors that robustly transfer from preclinical models to human tumors. TRANSACT performs favorably compared to four competing approaches, including two deep learning approaches, on a set of 23 drug prediction challenges on The Cancer Genome Atlas and 226 metastatic tumors from the Hartwig Medical Foundation. We demonstrate that response predictions deliver a robust performance for a number of therapies of high clinical importance: platinum-based chemotherapies, gemcitabine, and paclitaxel. In contrast to other approaches, we demonstrate the interpretability of the TRANSACT predictors by correctly identifying known biomarkers of targeted therapies, and we propose potential mechanisms that mediate the resistance to two chemotherapeutic agents.
引用
收藏
页数:12
相关论文
共 99 条
[1]   iTOP: inferring the topology of omics data [J].
Aben, Nanne ;
Westerhuis, Johan A. ;
Song, Yipeng ;
Kiers, Henk A. L. ;
Michaut, Magali ;
Smilde, Age K. ;
Wessels, Lodewyk F. A. .
BIOINFORMATICS, 2018, 34 (17) :988-996
[2]   TANDEM: a two-stage approach to maximize interpretability of drug response models based on multiple molecular data types [J].
Aben, Nanne ;
Vis, Daniel J. ;
Michaut, Magali ;
Wesseis, Lodewyk F. A. .
BIOINFORMATICS, 2016, 32 (17) :413-420
[3]   Machine learning and feature selection for drug response prediction in precision oncology applications [J].
Ali M. ;
Aittokallio T. .
Biophysical Reviews, 2019, 11 (1) :31-39
[4]   Drug response prediction by inferring pathway-response associations with kernelized Bayesian matrix factorization [J].
Ammad-ud-din, Muhammad ;
Khan, Suleiman A. ;
Malani, Disha ;
Murumagi, Astrid ;
Kallioniemi, Olli ;
Aittokallio, Tero ;
Kaski, Samuel .
BIOINFORMATICS, 2016, 32 (17) :455-463
[5]  
[Anonymous], 2008, Transfer learning via dimensionality reduction
[6]   THEORY OF REPRODUCING KERNELS [J].
ARONSZAJN, N .
TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY, 1950, 68 (MAY) :337-404
[7]   Akt-dependent nuclear localization of Y-box-binding protein 1 in acquisition of malignant characteristics by human ovarian cancer cells [J].
Basaki, Y. ;
Hosoi, F. ;
Oda, Y. ;
Fotovati, A. ;
Maruyama, Y. ;
Oie, S. ;
Ono, M. ;
Izumi, H. ;
Kohno, K. ;
Sakai, K. ;
Shimoyama, T. ;
Nishio, K. ;
Kuwano, M. .
ONCOGENE, 2007, 26 (19) :2736-2746
[8]  
Behdenna A, 2020, bioRxiv, DOI [10.1101/2020.03.17.995431, 10.1101/2020.03.17.995431, DOI 10.1101/2020.03.17.995431]
[9]   A theory of learning from different domains [J].
Ben-David, Shai ;
Blitzer, John ;
Crammer, Koby ;
Kulesza, Alex ;
Pereira, Fernando ;
Vaughan, Jennifer Wortman .
MACHINE LEARNING, 2010, 79 (1-2) :151-175
[10]   Genetic and transcriptional evolution alters cancer cell line drug response [J].
Ben-David, Uri ;
Siranosian, Benjamin ;
Ha, Gavin ;
Tang, Helen ;
Oren, Yaara ;
Hinohara, Kunihiko ;
Strathdee, Craig A. ;
Dempster, Joshua ;
Lyons, Nicholas J. ;
Burns, Robert ;
Nag, Anwesha ;
Kugener, Guillaume ;
Cimini, Beth ;
Tsvetkov, Peter ;
Maruvka, Yosef E. ;
O'Rourke, Ryan ;
Garrity, Anthony ;
Tubelli, Andrew A. ;
Bandopadhayay, Pratiti ;
Tsherniak, Aviad ;
Vazquez, Francisca ;
Wong, Bang ;
Birger, Chet ;
Ghandi, Mahmoud ;
Thorner, Aaron R. ;
Bittker, Joshua A. ;
Meyerson, Matthew ;
Getz, Gad ;
Beroukhim, Rameen ;
Golub, Todd R. .
NATURE, 2018, 560 (7718) :325-+