Inhalt

•  
   List of Abbreviations
•  
   Notation Tables
•  
   1 Introduction
•  
  2 Background
•  
   2.1 Probabilistic Models and Optimization
•  
   2.1.1 Mixed Integer Linear Programming
•  
   2.1.2 Markov Chain Monte Carlo
•  
   2.1.3 Model Selection
•  
  2.2 Biological and Technical Background
•  
   2.2.1 Cancer and Genetic Mutations
•  
   2.2.2 Next-Generation Sequencing Techniques
•  
   2.2.3 Detecting Somatic Mutations
•  
  2.3 Subclonal Reconstruction of Cancer Samples
•  
   2.3.1 Clonal Evolution Theory and Intratumor Heterogeneity
•  
   2.3.2 Formalized Problem Description
•  
   2.3.3 Subclonal Reconstruction Concepts and Methods
•  
   3 A New Lineage-Based Subclonal Reconstruction Model
•  
   3.1 The Likelihood Function
•  
   3.2 Model Components and Rules
•  
   3.2.1 Inferred Lineage Frequencies
•  
   3.2.2 Inferred Lineage Relationships
•  
   3.2.3 Copy Number Aberration Assignment
•  
   3.2.4 Simple Somatic Mutation Assignment
•  
  3.3 Optimization with Mixed Integer Linear Programming
•  
   3.3.1 Objective Function and Basic Mixed Integer Linear Program
•  
   3.3.2 Variables and Constraints for Lineage Frequencies
•  
   3.3.3 Variables and Constraints for Lineage Relationships
•  
   3.3.4 Variables and Constraints for Copy Number Aberrations
•  
   3.3.5 Variables and Constraints for Simple Somatic Mutations
•  
   3.3.6 Reducing the Number of Variables and Constraints
•  
   3.4 Optimization Complexity
•  
   3.5 Determining the Number of Lineages
•  
  4 Dealing with Ambiguity
•  
   4.1 Defining Ambiguity
•  
   4.2 Handling Ambiguity
•  
   4.2.1 Finding Present Ancestor-Descendant Relationships Necessary because of Likelihood Influence
•  
   4.2.2 Updating Lineage Relationships
•  
   4.2.3 Unphasing Simple Somatic Mutations
•  
   4.2.4 Identifying Absent Ancestor-Descendant Relationships Necessary because of Crossing Rule and Mutation Assignment
•  
   4.2.5 Identifying Present Ancestor-Descendant Relationships Necessary because of Sum Rule
•  
   4.2.6 Identifying Absent Ancestor-Descendant Relationships Necessary because of Sum Rule
•  
   4.3 Lineage-Based versus Population-Based Subclonal Reconstruction
•  
  5 Analyzing Onctopus' Performance
•  
   5.1 Implementation
•  
   5.2 Data Simulation
•  
   5.3 Evaluation Metrics
•  
   5.4 Optimality, Run Time and Memory Usage
•  
   5.4.1 General Experiment
•  
   5.4.2 Increasing Run Time
•  
   5.4.3 Conclusion
•  
  5.5 Clustering Simple Somatic Mutations
•  
   5.5.1 Clustering Algorithms and Cluster Numbers
•  
   5.5.2 Building Subclonal Reconstructions with Clustered Simple Somatic Mutations
•  
   5.6 Fixing Copy Number Aberrations
•  
  5.7 Fixing Lineage Frequencies
•  
   5.7.1 Performance with Correct Lineage Frequencies
•  
   5.7.2 Inference of Lineage Frequencies Depending on the Number of Simple Somatic Mutations
•  
   5.7.3 Performance with Inferred Lineage Frequencies
•  
   5.8 Approximating Variant Allele Frequencies in Mixed Integer Linear Program
•  
  6 Results and Evaluation
•  
   6.1 Evaluation Metrics
•  
   6.2 Results on Simulated Data
•  
   6.2.1 Data Simulation
•  
   6.2.2 Inferring Subclonal Reconstructions
•  
   6.2.3 Results
•  
   6.2.4 Discussion
•  
  6.3 Results on a Breast Cancer Dataset
•  
   6.3.1 Data Description
•  
   6.3.2 Inferring Subclonal Reconstructions
•  
   6.3.3 Results and Discussion
•  
   7 Conclusion and Outlook
•  
   Bibliography
•  
   A Onctopus Software
•  
  B Data Simulation
•  
   B.1 Data Simulation
•  
   B.2 Simulated Datasets for Analyzing Optimality, Run Time and Memory Usage
•  
  B.3 Simulated Datasets for Simple Somatic Mutation Clustering Analysis
•  
   B.3.1 Clustering Algorithms and Cluster Numbers
•  
   B.3.2 Building Subclonal Reconstructions with Clustered Simple Somatic Mutations
•  
   B.4 Simulated Datasets for Fixing Copy Number Aberration Analysis
•  
  B.5 Simulated Datasets for Fixing Lineage Frequencies Analysis
•  
   B.5.1 Simulated Datasets for Inference of Lineage Frequencies Depending on the Number of Simple Somatic Mutations
•  
   B.5.2 Simulated Datasets for Analysis of Performance with Inferred Lineage Frequencies
•  
   B.6 Simulated Datasets for Analysis of Approximating Variant Allele Frequencies in Mixed Integer Linear Program
•  
   B.7 Simulated Datasets for Comparison between Onctopus, PhyloWGS and Canopy