COS598G: Topics in Computational Biology: Spring 2017

Outline of Topics (Tentative)

Our guiding application will be cancer genomics and immunogenomics.  However, computational approaches that we study have a variety of applications in computational biology and related fields.

Topic 1: Phylogenetic tree reconstruction [3.5 weeks.  4 lectures; 3 discussion]

1. Review/introduction: Algorithms for phylogenetic tree reconstruction 

• Character­s vs. distance
• Small phylogeny problem vs. large phylogeny problem
• Maximum parsimony (Sankoff’s algorithm)
• Maximum likelihood (Felsenstein algorithm)
• Large phylogeny: searching through tree space

2. Important special cases of maximum parsimony

• Perfect phylogeny (Infinite sites assumption)
• Multi-state perfect phylogeny (Infinite alleles assumption)
• Cladistic multi-state perfect phylogeny vs. two-state perfect phylogeny
• Dollo parsimony

3. Sequencing of cancer genomes
4. Papers and discussion

• Sequencing mixtures; Bulk-sequencing of tumors

• Review: Clonal Heterogeneity and Tumor Evolution: Past, Present, and the Future
• CompBio: TrAp: a tree approach for fingerprinting subclonal tumor composition
• CompBio: Reconstruction of clonal trees and tumor composition from multi-sample sequencing data

• Perfect phylogeny with errors; single-cell sequencing

• Review:  Van Loo P, Voet T. Single cell analysis of cancer genomes. Curr Opinion Genet Dev. 2014; 24:82–91
• Algorithms: Exact ILP Solutions for Phylogenetic Minimum Flip Problems.  Chimani, Rahmann, and Bocker
• CompBio: Tree inference for single-cell data.  Jahn, Kuipers, Beerenwinkel.  Genome Biology (2016)
• CompBio: OncoNEM: inferring tumor evolution from single-cell sequencing data. Ross and Markowetz.  Genome Biology (2016)
• CompBio: Using single cell sequencing data to model the evolutionary history of a tumor.  Kim and Simon.  BMC Bioinformatics (2014)

Copy number aberrations and multi-state phylogeny

• CompBio: Assessing intratumor heterogeneity and tracking longitudinal and spatial clonal evolutionary history by next-generation sequencing

Topic 2: Population Genetics models [3 weeks; 3 lectures, 3 discussion]

1. Review/introduction: Population genetics

1. Forward-time models: Wright-Fisher and Moran Model
2. Backward-time model and coalescent.

2. Methods to detect selection
3. Exponentially growing populations
4. Papers and discussion

1. Review:
2. Biology/Combio:  Cumulative Haploinsufficiency and Triplosensitivity Drive Aneuploidy Patterns and Shape the Cancer Genome
3. CompBio: Identification of neutral tumor evolution across cancer types
4. Math/CompBio: POPULATION GENETICS OF NEUTRAL MUTATIONS IN EXPONENTIALLY GROWING CANCER CELL POPULATIONS
5. CompBio: Quantification of subclonal selection in cancer from bulk sequencing data
6. CompBio (no cancer): Learning natural selection from the site frequency spectrum
7. CompBio (no cancer): Predicting Carriers of Ongoing Selective Sweeps without Knowledge of the Favored Allele

Topic 3: Network analysis of biological data  [3.5 weeks.  4 lectures; 3 discussion]

1. Random walks and diffusion on graphs

1. Lazy random walk: symmetric versus asymmetric
2. Random walk with restart, PageRank
3. Heat equation and heat kernels

2. Graph partitioning        

1. Graph conductance
2. Community detection and Modularity

3. Papers and discussion

1. Review: Network propagation: a universal amplifier of genetic associations
2. Review: Understanding Genotype-Phenotype Effects in Cancer via Network Approaches
3. Review: Computational Solutions for Omics Data
4. Review: Network Approaches to Complex Disease
5. CompBio: Walking the Interactome for Prioritization of Candidate Disease Genes
6. CompBio: Network-Based Integration of Disparate Omic Data To Identify "Silent Players" in Cancer
7. CompBio: Discovering causal pathways linking genomic events to transcriptional states using Tied Diffusion Through Interacting Events (TieDIE)
8. CompBio: Exploiting ontology graph for predicting sparsely annotated gene function
9. CompBio: Compact Integration of Multi-Network Topology for Functional Analysis of Genes
10. Math/Stats: Near-optimal Anomaly Detection in Graphs using Lova ́sz Extended Scan Statistic
11. CS: Vertex Neighborhoods, Low Conductance Cuts, and Good Seeds for Local Community Methods.

Topic 4: Immunogenomics [2 weeks.2 lectures; 2 discussion]

1. Introduction: Components of the immune system:  B-cells; T-cells; MHC
2. Sequencing of B-cell and T-cell repertoires
3. Modeling of clonal expansions of B-cells
4. Epitope prediction.
5. Papers and discussion

1. Review: The promise and challenge of high-throughput sequencing of the antibody repertoire.  Nature Biotechnology (2014)
2. Review: Solving Immunology
3. Review: The Diversity and Molecular Evolution of B-Cell Receptors during Infection.  
4. CompBio: Quantifying selection in high-throughput Immunoglobulin sequencing data sets.   Kleinstein NAHR (2012)
5. CompBio: Identifying T Cell Receptors from High-Throughput Sequencing: Dealing with Promiscuity in TCRα and TCRβ Pairing.   PLOS Computational Biology (2017)
6. CompBio: Single-cell TCRseq: paired recovery of entire T-cell alpha and beta chain transcripts in T-cell receptors from single-cell RNAseq
7. CompBio: Gapped sequence alignment using artificial neural networks: application to the MHC class I system. Bioinformatics (2016)
8. CompBio: MHCflurry Hammerbacher (2016): http://biorxiv.org/content/biorxiv/early/2016/05/23/054775.full.pdf