CS8395 Special Topics in Computational Biology

🧭 Description

This graduate seminar examines how computational methods are designed, assessed, and advanced to tackle modern biological questions, with emphasis on applying machine learning and AI to model, integrate, and interpret single-cell and spatial transcriptomic data. In particular the course explores how choices in data representation, model specification, and algorithm design shape what can be inferred from such datasets.

Through discussions and student presentations of recent work from both conferences (e.g., ISMB, RECOMB, NeurIPS, MLCB etc.) and journals (Nature Methods, Nature Biotechnology etc.), students will learn how to critically evaluate existing models, identify the underlying assumptions and limitations, and develop their own ideas. The course also includes hands-on mini-projects in which students will design, implement, and assess a computational method inspired by the papers discussed in class.

💡 Beyond running analyses, we focus on how to think like a method developer — turning ideas into models that connect computation and biology.

Logistics

• Course Code: CS8395-04
• Term: Spring 2026
• Class Times: Tuesdays & Thursdays, 02:45 PM - 04:00 PM
• Location: Olin Hall 131

🧬 Computational biology in a Nutshell

🎯 Learning Goals

By the end of this course, students will be able to:

• Define computational biology and its scope across computer science and life sciences
• Understand major biological data types — single-cell, spatial, proteomic, genomic, and multi-omics
• Formulate computational methods to address biological problems
• Evaluate and reconstruct mathematical models from research literature
• Appreciate the role of simulation and benchmarking in method development
• Translate mathematical models into efficient, reproducible implementations

🗓️ Lecture Outline

🧩 Core Topics (Summary)

1. Foundations
   • What is computational biology?
   • Designing methods vs running tools
2. Data Representation
   • Single-cell, spatial, and multi-omics data
   • Structuring biological datasets
3. Exploratory Analysis
   • Visualization, dimensionality reduction, clustering
   • Feature selection in multi-omics contexts
4. Problem Formulation
   • Translating biological questions into computational problems
   • Case studies from recent literature
5. Modeling & Software
   • Statistical, probabilistic, and ML frameworks
   • Coding (Python, R, C++, Rust), efficiency, and reproducibility
6. Simulation & Benchmarking
   • Role of simulations and benchmark datasets
   • Designing fair, reproducible evaluations

🏗️ Course Structure

Lectures/Discussions: Weekly topics introduced through short lectures and literature discussions (back-to-back or two days of the same week).

Readings: Research papers, book chapters, and review articles.

Assignments: Focused problem sets and small coding projects.

Final Project: Students will select a computational biology problem, survey existing literature, and either (1) reconstruct/improve an existing method or (2) propose/implement a novel approach.

👥 Evaluation (Groups)

📚 Suggested References

• Compeau & Pevzner, BIOINFORMATICS ALGORITHMS - An Active Learning Approach
• John Tukey, Exploratory Data Analysis
• Pevzner & Shamir, Computational and Systems Biology
• Blum, Hopcroft & Kannan, Foundations of Data Science
• Single-cell Best Practices
• Selected research papers (to be assigned weekly)

🌎 Similar Courses at Other Universities

• Princeton - COS 598E: Methods in Computational Biology
• Virginia Tech - CS6222: Introduction to Bioinformatics
• Cogniterra - Computational Biology Course
• UPenn - GCB 5370: Advanced Topics in Computational Biology