Quantitative and Computational Biology

--

--

--

--

This interdisciplinary course provides a broad overview of computational and experimental approaches to decipher genomes and characterize molecular systems. We focus on methods for analyzing "omics" data, such as genome and protein sequences, gene expression, proteomics and molecular interaction networks. We cover algorithms used in computational biology, key statistical concepts (e.g., basic probability distributions, significance testing, multiple testing correction, performance evaluation), and machine learning methods which have been applied to biological problems (e.g., classification techniques, hidden Markov models, clustering).

Analysis of recent work on quantitative, theoretically grounded approaches to the phenomena of life. Topics rotate from year to year, spanning all levels of biological organization, including (as examples) initial events in photosynthesis, early embryonic development, evolution of protein families, coding and computation in the brain, collective behavior in animal groups. Assumes knowledge of relevant physics and applicable mathematics at advanced undergraduate level, with tutorials on more advanced topics. Combination of lectures with student discussion of recent and classic papers.

This course establishes a foundation in applied statistics and data science for those interested in pursuing data-driven research. The course may involve examples from any area of science, but it places a special emphasis on modern biological problems and data sets. Topics may include data wrangling, exploration and visualization, statistical programming, likelihood based inference, Bayesian inference, bootstrap, EM algorithm, regularization, statistical modeling, principal components analysis, multiple hypothesis testing, and causality. The statistical programming language R is extensively used to explore methods and analyze data.

This course focuses on computational methods in cryo-EM, including three-dimensional ab-initio modelling, structure refinement, resolving structural variability of heterogeneous populations, particle picking, model validation, and resolution determination. Special emphasis is given to methods that play a significant role in many other data science applications. These comprise of key elements of statistical inference, image processing, and linear and non-linear dimensionality reduction. The software packages RELION and ASPIRE are routinely used for class demonstration on both simulated and publicly available experimental datasets.

A chemically and quantitatively rigorous treatment of metabolism and protein synthesis, with a focus on modern advances and techniques. Topics include metabolic pathways and their regulation; metabolite and flux measurement; mathematical modeling of metabolism; amino acid, peptide and protein chemistry; protein engineering and selected applications thereof.