Syllabus

Course Description 

The availability of large-scale human behavior, biological, clinical, and public health data and new developments in data science have enabled discovery of novel knowledge beyond human intuition. This course introduces fundamental concepts and computational techniques of systems modeling to extract biomedical insights from big heterogeneous data. Students will learn characteristics of large-scale multimodal health data, best-practices for data harmonization and fusion, state-of-the-art AI and statistical methods of integrative analysis, and approaches to model interpretation and causality. This class uses a combination of formal lectures and computer lab exercises. Real-world examples will be used to demonstrate the algorithms, implementation, and application of different techniques. Students will have hands-on practice of systems modeling to solve precision nutrition problems.

Prerequisites 

Given the many and unique backgrounds of trainees in this project, we will require all students to complete leveling courses on basic nutrition science, statistics, and machine learning before enrolling in this course. 

This course will use Python to demonstrate algorithms and applications. Therefore, students are expected to have prior programming experience in at least one language (preferable Python, R, or Matlab). 

This course will use cloud-computing resources and ASU high-performance computing clusters. Therefore, prior experience in UNIX/LINUX environment is desirable.

Learning Outcomes 

Students will gain theoretical and practical knowledge of systems modeling and applications to precision health, including (i) survey of big health data, (ii) recent developments in the field, (iii) full life-cycle of systems modeling projects, (iii) understanding of core algorithms, (iv) ability to perform systems modeling and evaluate the results, (v) challenges and future research in systems modeling that may influence precision health. At the end of the course, students should be able to decide if a non-trivial problem can be solved effectively using systems modeling, and if so, design a prototype system to solve the problem.

Real data will be utilized from ASU faculty projects and publicly available databases for class lectures and assignments with the potential to publish analyses completed during the course.

Expectations

 Students will form groups of 2-3, and complete a final project of integrating and analyzing real data selected by the group members, discussed and approved by the instructor. Student teams will apply knowledge and skills acquired in the class to conduct the analysis and to interpret the results. The process – including protocols, results and discussions – will be posted on GitHub, reviewed by your peers and the instructor. A write-up by each group describing the project is due at the end of the course, and evaluated by the instructor. It should cover basic elements of a research article, such as introduction, methods, results, and discussion.

Required Materials

A computer (laptop, desktop, any operating system) will be required for non-computer lab work. Discuss with the instructor in advance if you do not have access to a suitable computer. All students will receive temporary access to the ASU supercomputer to complete computationally-intensive tasks.

Credits: 3

Course Format: In-person lectures & computer lab exercises

Course Meeting Times: 2 sessions / week, 1.5 hours / session

Schedule of Topics:

Week 1: What is systems science? [Liu, Whisner]

• Lecture: Systems science across the translational research spectrum

• Lecture: Systems used to study nutrition and precision health

• Assignment - Discussion Board: Student introductions and reflection on systems science and translational research

Week 2: Big data for precision nutrition studies [Liu, Sears, Dinu]

• Lecture: Multimodality and heterogeneity of nutrition and health data

• Lecture: Data interconnectivity

Week 3: Big data for precision nutrition studies [Liu, Sears, Dinu]

• Lecture: Selecting datasets and databases for impactful research questions

• Computer Lab: Explore and discuss data repositories

• Assignment: Identify 2-3 datasets or databases that you find interesting and could be integrated to answer systems-level research questions. Provide three potential research questions you could ask using the selected datasets.

Week 4: Aligning analytical systems to biological and theoretical frameworks [Whisner]

• Lecture: Life Course theory for modeling lifelong exposures and dynamic life stages

• Lecture: Metabolic networks and systems

• Assignment - Discussion Board: What challenges are you facing in the course so far?

Week 5: Aligning analytical systems to biological and theoretical frameworks [Buman, Rivera, Ojinnaka]

• Lecture: Socio ecological models (e.g., social cognitive theory), social determinants of health and social networks analysis

• Lecture: Dynamics of complex adaptive systems, control systems, and population-based modeling

• Assignment: Submit finalized research question for final project

Week 6: Data harmonization, fusion & augmentation [Turaga, Kocher]

• Lecture/Computer Lab: Synthesis and alignment of data

• Lecture: Early/intermediate/late fusion

• Assignment: Create a 10 minute presentation discussing a biological or theoretical framework that could help guide a research question using the datasets selected in Week 3.

Week 7: Data harmonization, fusion & augmentation [Buman, Turaga, Kocher]

• Lecture/Computer Lab: Enrichment with annotations, ontologies, etc.

• Lecture: AI-ready data sets, etc.

• Assignment: Create and implement a plan for synthesizing, aligning and fusing data for the final project.

Week 8: Integrative analysis incorporating complex relationships [Scotch, Ghasemzadeh, Runger]

• Lecture/Computer Lab: Modeling network structure

• Lecture/Computer Lab: Modeling spatial distributions

• Assignment: Create and implement a plan for enriching final project data with annotations and ontologies

Week 9: Integrative analysis incorporating complex relationships [Scotch, Ghasemzadeh, Runger]

• Lecture/Computer Lab: Modeling longitudinal data

• Lecture/Computer Lab: Modeling mediations

• Assignment: Data modeling practice set 1

Week 10: Integrative analysis incorporating complex relationships [Runger, Chen]

• Modeling prior knowledge of nutrition science

• Causal inference

• Assignment: (1) Data modeling practice set 2 and (2) Submit draft of analysis plan for final project

Week 11: Implementing systems-based analyses [Liu, Whisner]

• Computer Lab: Instructor-supported analytics day for final project

• Computer Lab: Instructor-supported analytics day for final project

• Assignment - Discussion Board: Course progress reflection and self-assessment of skill development

Week 12: Analysis Interpretation [Liu, Whisner]

• Lecture/Computer Lab: Evaluation of modeling results, e.g., predictions, clusters, etc.

• Lecture/Computer Lab: Understanding model parameters

• Assignment - Discussion Board: Discuss preliminary findings from Week 11 analyses

Week 13: Implementing systems-based analyses [Liu, Whisner]

• Computer Lab: Instructor-supported analytics day for final project

• Computer Lab: Instructor-supported analytics day for final project

• Assignment - Discussion Board: Reflect on analysis plan progress, changes, and needs

Week 14: Analysis Interpretation [Liu, Whisner]

• Lecture: Formulation of new hypotheses

• Lecture/Computer Lab: Fine-tuning models to enhance interpretation and nutrition/population relevance

Week 15: Final Project Completion [Liu, Whisner]

• Lecture: Confirmatory testing

• Computer Lab: Instructor-supported analytics day for final project

• Assignment: Submit final project background, methods, results and interpretation