From Regression to LLMs: Introduction to Machine Learning for Computational Biology
This course will provide a rigorous treatment of computational statistics and machine learning methods used to analyze big biological data types. Analysis methods covered will include: generalized linear models, support vector machines, regularized linear models, kernel methods, random forests, neural networks, large language models. While the course will be focused on analysis methods and connections among methods, applications making use of specific big biological data types will be covered. An understanding of method limitations will be prioritized, as well as how to critically assess when a desired conclusion can be justified. Methods will be implemented in the computer lab in Python, where some previous exposure to programming will be assumed.
Topics
- Theory & Practice - Students will be able to explain the formal mathematical framework for each of the following methods, as well as implement a version of each in Python: generalized linear models, support vector machines, regularized linear models, kernel methods, random forests, neural networks, large language models.
- Data Analysis - Students will apply each of the following methods to both simulated and real biological data, including genomic, image, and health record data: generalized linear models, support vector machines, regularized linear models, kernel methods, random forests, neural networks, large language models.
- Communicating & Arguing with Data - Students will plot, present, and interpret in writing the outcomes from applying each of the following methods to real genomic, image, or health record data: generalized linear models, support vector machines, regularized linear models, kernel methods, random forests, neural networks, large language models.
- Data Interpretation and Critique - Students will be able to rigorously explain the limitations, and the implications of these limitations, when using the following methods to draw specific and alternative biological conclusions from real data: generalized linear models, support vector machines, regularized linear models, kernel methods, random forests, neural networks, large language models.
- Rigorously define a random vector and a sample (Assessment: student can write the exact definition of random vector and sample from memory).
- Define p-value and calculate a p-value for linear and logistic regression models (Assessment: student can write the exact definition of a p-value from memory and can implement R or Python code to calculate a p-value when applying a linear or logistic regression).
- Implement a cross-validation assessment of model fit for a Support Vector Machine (Assessment: student can implement R or Python code to perform a cross-validation performance assessment for a Support Vector Machine model when provided with necessary parameters).
- Train a neural network to predict an outcome variable (Assessment: student can implement R or Python code to train a neural network on provided data).
Official roster learning outcomes
Sources
- Course catalogue ↗2026-2027
- Catalogue entry ↗
Catalogue variants
- Computational Statistics & Machine Learning in Bio · 4.0 · Graduate