Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/97486
Title: Predicción de respuesta a fármacos quimioterapéuticos a partir de datos genómicos
Author: Esteban Lasso, Alfonso
Tutor: Tejero, Héctor  
Others: luna, Jeroni  
Abstract: Cancer is one of the leading causes of death in the world. Despite the great advances made in recent decades, in many cases the tumors do not respond to standard treatment or develop resistance during treatment. In order to facilitate future personalized treatments, a series of high-performance genomic technologies are being developed, including Machine Learning algorithms for predicting drug response. With this in mind, I have applied a series of Machine Learning algorithms with different combinations of adjustment customization in order to develop models capable of predicting, as accurately as possible, the response to drugs registered at the GDSC and DepMap Broad Institute. These algorithms have been trained and adjusted for 4 random drugs to see the accuracy of prediction on their corresponding expression data based on the AUC values obtained by these two institutions with favorable results and obtaining the 2 best combinations tested: The drug that is best predicted is Erlotinib with 87.64% accuracy, accurate predictions in a 60/20/20 data partition, by means of Random Search in Random Forest when the Auc values are discretised. When the predictors have been trained as continuous values, the best R2 value obtained was 1 corresponding to the response prediction for Erlotinib, Rapamycin and Sunitinib drugs using the adjusted Ridge regularization model with hyperparameters selected by RandomSearch and with an 80/20 partition in 10 en validación cruzada folds.
Keywords: algorithms
prediction
cancer
Document type: info:eu-repo/semantics/masterThesis
Issue Date: 6-Jun-2019
Publication license: http://creativecommons.org/licenses/by-nc-nd/3.0/es  
Appears in Collections:Trabajos finales de carrera, trabajos de investigación, etc.

Files in This Item:
File Description SizeFormat 
Memoria Final TFM.zip137,32 MBUnknownView/Open

presentación TFM.mp4

32,46 MBMP4View/Open
aestebanlasTFM0619memoria.pdfMemoria del TFM7,71 MBAdobe PDFThumbnail
View/Open