Machine Learning-based Prediction of the Three Drugs Combination Synergy Öffentlichkeit

Abstract

Background: The development of novel drug combination therapy is one of the hot research areas to treat complex diseases. However, current drug combination research still using relatively inefficient experimental screening methods. Though some researchers have already developed in-silico tools for predicting drugs combination synergy, they only focused on two drugs combinations. In this study, we tried to predict three drugs combination synergy using machine learning approaches, and compared models’ performance by different algorithms and datasets. It’s the first study to explore predicting the synergy response of three-drug combinations.

Methods: In our datasets, we included three drugs combination synergy responses, drug dosages for 560 combination levels, gene expression and mutation data for 13 cell lines. We mainly used tree models for feature selection. We tried modern tree models, Support Vector Machine, and Deep Neural Network for model selection. We also explored whether include extra cell lineage information, gene expression and mutation data in the dataset would improve the model’s performance. Furthermore, we developed a novel LightGBM-based vote model and compared its performance with other models.

Results: Adding extra cell lineage or gene expression / mutation data would improve the model’s performance. The LightGBM model showed best performance among traditional models, while our novel vote model beat it and even achieved better results according to the cross-validation results.

Table of Contents

1 INTRODUCTION 1

2 MATERIALS AND METHODS   5

2.1 DATASETS      5

2.1.1 Three drugs combination synergy experiments     5

2.1.2 Genomic features and Simple indicators for cell line types   6

2.2 EXPERIMENTS 7

2.2.1 Feature selection 7

2.2.2 Machine learning methods comparison  7

2.2.3 Dataset comparison   11

2.2.4 Threshold comparison 11

2.2.4 A vote model for three drugs combination    12

2.3 LEAVE ONE CELL LINE OUT CROSS VALIDATION     13

3 RESULTS    14

3.1 SYNERGY SCORES AND DATASETS    14

3.2 ALGORITHMS COMPARISON      15

3.3 DATASET COMPARISON      19

3.4 THRESHOLDS COMPARISON      19

4 DISCUSSION     21

ACKNOWLEDGMENT  24

BIBLIOGRAPHY    24

APPENDIX    27

About this Master's Thesis

Rights statement

• Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.

School	Rollins School of Public Health
Department	Biostatistics
Subfield / Discipline	Biostatistics - MPH & MSPH
Degree	M.S.P.H.
Submission	Master's Thesis
Language	English
Research Field	Biology, Bioinformatics Biology, Biostatistics
Stichwort	Drug combination synergy machine learning LightGBM Vote model
Committee Chair / Thesis Advisor	Qin, Zhaohui (Steve), Emory University
Committee Members	Dreaden, Erik C., Emory University

Zuletzt geändert

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	Machine Learning-based Prediction of the Three Drugs Combination Synergy ()	2021-04-24 10:11:30 -0400	Press to Select an action Download

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions