CS690: Computational Genomics

About the Course:

Computational genomics is a novel and very active application field of computer science where biological mechanisms are deciphered from genome sequencing data using computational and statistical analyses. In the past fifteen years, an explosion of genomic data (from human and several other organisms) has revolutionized a number of subfields of biology – cell and molecular biology, developmental biology, disease biology and so on. Computer science plays a central role in genomics – from sequencing and assembling of DNA and RNA sequences to analyzing genomes (or transcriptomes) for elucidating diverse biological mechanisms through innovations in machine learning, data structure and algorithms. In this course, you will be introduced to some of the most seminal machine learning and algorithmic approaches for sequence analysis as well as the most recent advances in the field. The course will be structured as a combination of lectures and discussion of recent publications in the field. The lectures will introduce the topics and seminal algorithms followed by research paper discussions on advanced and most recent developments. 

Tentative Topics:

• Transcriptomics: Gene expression analysis, normalization, differential expression analysis, clustering 

• Single-cell omics: Dimension reduction algorithms, Generative processes, Manifold learning, Cell type clustering and annotation, Trajectory inference, Data integration, Cell cell communication inference, Regulatory networks

• Spatial Transcriptomics: Cell type deconvolution – graphical models, Clustering of Spatial Data 

• Genome-Scale Index Structures: Suffix array, Suffix tree, Burrows-Wheeler transform (BWT), BWT index and applications 

• Genome-Scale Short Read Alignment: Dynamic programming along suffix tree paths, short read alignment algorithms

• Variant Calling Algorithms: Probabilistic dynamic programming

• Cancer Genomics: Probabilistic graphical models for tumor heterogeneity analysis, somatic mutation detection algorithms, driver mutation detection, matrix factorization problems in cancer genomics

• Phylogenetics: Markov models of molecular evolution, character-based phylogeny algorithms – maximum parsimony, maximum likelihood and Bayesian inference