ECE 471

Extracting insights from heterogeneous datasets to support decision-making is fundamental to modern applications. This course teaches students to engineer analysis workflows that use feature engineering, longitudinal machine learning methods, and validation to derive real-world insights from data. Students gain hands-on experience through lectures and labs and via three projects involving large-scale real-world data from domains such as autonomous-vehicles, healthcare and trust. While each workflow is end-to-end, students will delve deeper into methods as the course progresses. Course Information: 3 undergraduate hours. 4 graduate hours. Prerequisite: Basic probability and basic computer programming skills are essential. ECE 313 or CS 361. Prior exposure to basics of scripting languages (such as Python), knowledge of operating systems (e.g., ECE 391, or an equivalent course) is beneficial.

The goal of this course is to highlight elementary design principles of biological systems. Many of the underlying principles that govern the biochemical interactions within a cell can be related to networks consisting of basic building-block circuits with multiple inputs/output, feedback and feedforward etc. This course draws on control theory and simple biology to provide a mathematical framework to understand these biological networks. The course is intended for advanced undergraduates or graduate students.

Grade is determined by a mid-term (40%), final exam (40%), and homeworks + class participation in analysis of current research (20%).

• Introduction to molecular/cell biology
• A review of network concepts: properties and modeling of feedback/feed-forward systems
• Introduction to molecular biology (molecular recognition, proteins, DNA, repressors/promoters/ Hill functions)
• transcription networks (timescales, introduction to gene regulation)
• Transcription networks revisited (multi-d input functions, dynamic response in gene regulation)
• Autoregulation (AR) (negative for fast response time and robust stable production in gene circuits)
• Autoregulation (positive AR slows response and leads to bi-stability
• Feed-forward Loop (FFL) Network Motif (Dynamics of coherent FFL with AND logic)
• FFL is a sign-sensitive delay element
• Incoherent FFL (dynamics—pulse generator; response acceleration)
• Network Motifs in development transcription networks (positive feedback loops for making decisions; regulating feedback; developmental timing; interlocked feed-forward loops in B. subtilis)
• Information processing using Multi-layer perceptrons.
• Network Motifs in neuronal networks (An example: C. elegans)
• Network Motifs: negative feedback and oscillator motifs
• Protein circuits (a review of protein biochemistry)
• Protein circuits (an example: bacterial chemotaxis in E. coli)  
• Two models for Adaptation: 1. Robust and 2. Fine-tuned
• The Robust Adaptation (Barkai-Leibler) and Integral Feedback.
• Linearization of nonlinear systems--linear system response
• Stability— Routh criterion, Nyquist criterion, root locus techniques,
• Circadian rhythms—how to build an oscillator; represillator
• Buzzers, Toggles, sniffers, and oscillators
• Kinetic Proofreading (proofreading the genetic code to reduce error rates of molecular recognition)
• Recognizing Self and Non-self by the Immune system
• Kinetic Proofing and T-cell recognition
• Gene Circuit Design I (optimal expression of a protein in a constant environment)
• Gene Circuit Design II(optimal regulation in a variable environment)
• The Savageau Demand Rule: e.g. the demand rule in E. coli
• Rules for gene regulation (based on minimal error load or selection repression; multiregulator systems)

Topics:

• Introduction to molecular/cell biology
• A review of network concepts: properties and modeling of feedback/feed-forward systems
• Introduction to molecular biology (molecular recognition, proteins, DNA, repressors/promoters/ Hill functions)
• transcription networks (timescales, introduction to gene regulation)
• Transcription networks revisited (multi-d input functions, dynamic response in gene regulation)
• Autoregulation (AR) (negative for fast response time and robust stable production in gene circuits)
• Autoregulation (positive AR slows response and leads to bi-stability
• Feed-forward Loop (FFL) Network Motif (Dynamics of coherent FFL with AND logic)
• FFL is a sign-sensitive delay element
• Incoherent FFL (dynamics—pulse generator; response acceleration)
• Network Motifs in development transcription networks (positive feedback loops for making decisions; regulating feedback; developmental timing; interlocked feed-forward loops in B. subtilis)
• Information processing using Multi-layer perceptrons.
• Network Motifs in neuronal networks (An example: C. elegans)
• Network Motifs: negative feedback and oscillator motifs
• Protein circuits (a review of protein biochemistry)
• Protein circuits (an example: bacterial chemotaxis in E. coli)  
• Two models for Adaptation: 1. Robust and 2. Fine-tuned
• The Robust Adaptation (Barkai-Leibler) and Integral Feedback.
• Linearization of nonlinear systems--linear system response
• Stability— Routh criterion, Nyquist criterion, root locus techniques,
• Circadian rhythms—how to build an oscillator; represillator
• Buzzers, Toggles, sniffers, and oscillators
• Kinetic Proofreading (proofreading the genetic code to reduce error rates of molecular recognition)
• Recognizing Self and Non-self by the Immune system
• Kinetic Proofing and T-cell recognition
• Gene Circuit Design I (optimal expression of a protein in a constant environment)
• Gene Circuit Design II(optimal regulation in a variable environment)
• The Savageau Demand Rule: e.g. the demand rule in E. coli
• Rules for gene regulation (based on minimal error load or selection repression; multiregulator systems)

Grade is determined by a mid-term (40%), final exam (40%), and homeworks + class participation in analysis of current research (20%).

2/13/2013