changed 4 years ago  
Linked with GitHub

Syllabus for GGG 298: Tools to support data-intensive research (Winter 2020/2021)

UC Davis

URL: http://bit.ly/2021-ggg-298-syllabus

Code of Conduct

Please abide by my lab's Code of Conduct in this course.

In particular, this is not an intellectual contest, and please realize that we all have plenty of things to learn.

Course sessions

GitHub site for class - redundant with links below, but more permanent!

Lab

What: hands-on computational work
When: Wed 9am-noon
Where: Zoom on Canvas

Discussion

What: read & discuss a paper.
When: Fri 12-1pm
Where: Zoom on Canvas

Homework

There will be 8 paragraph-length homeworks due, one each week on the reading; they'll be assigned a week in advance and due on Fridays at 11am.

Each week several (2-3) students will be asked to take a lead in preparing for the paper discussion on Fridays. Each student will only be asked to do this once. This counts as 3 homeworks instead of 1, for a total of 10.

Grading

The course is S/U, and only graded on homework; you need to hand in 7 of the 10 homeworks to pass.

Office hours

Titus will hold office hours online, from 2-4pm on Tu and Th at zoom.us/my/titusbrown/. Please sign up here in advance, or contact him via e-mail at ctbrown@ucdavis.edu: https://calendly.com/ctitusbrown/office-hours.

Please e-mail Shannon at sejoslin@ucdavis.edu to meet with her.

If neither of the listed office hours work and you have questions please email Shannon (sejoslin@ucdavis.edu) to setup a time to meet.

Instructors

C. Titus Brown (IOR) (ctbrown@ucdavis.edu), Shannon Joslin (sejoslin@ucdavis.edu).

Course description

This course will provide a practical introduction to common tools used in data-intensive research, including the UNIX shell, version control with git, RMarkdown, JupyterLab, and workflows with snakemake. The associated discussion section will connect the lab practicals to foundational concepts in data science, including repeatability/reproducibility, statistics, and publication ethics.

This course is open to all graduate students. No prior computational experience is required or assumed. There will be some minimal overlap with GGG 201(b) topics. All materials will be open to the community and freely available online.

Attendance and recordings

Lab sessions (on Wednesday) will be recorded and saved for a month, in Canvas. Please ask if you can't find one!

Discussion sessions (on Friday) will not be recorded.

Attendance is expected for both Wednesday and Friday. Please let Dr. Brown know if you can't attend.

In-class chat via Slack

We'll use Slack for sharing links and commands on both days. Please visit ucdavis.slack.com to create an account, and then join the slack channel #2021-ggg298.

Schedule of lab topics

Wednesdays, 9-noon

These will be lab practicals where we take a solid look at a given piece of technology.

From week 2 onwards, there will be ~20 minutes of videos to watch. Please watch them by 9:30am. At 9:30, the zoom will start and go 'til noon (with a 10 minute biobreak at 10:30). During this period we'll work through the day's lesson interactively.

You can view the lab topics here:

https://github.com/ngs-docs/2021-GGG298/

Paper discussions

Fridays, noon-1pm.

These will be discussion periods where we explore the concepts and questions introduced by various papers, articles, and blog posts. The topics will mostly be on data science and the practice of science.

First reading:

Our first paper is Trust but Verify: How to Leverage Policies, Workflows, and Infrastructure to Ensure Computational Reproducibility in Publication, by Willis and Stodden.

Please use this form to answer the following question by 11am on Friday, January 8th: Who decides if a paper is sufficiently reproducible?

In class on Friday we may discuss this question as well: How do you think "transparency" (as defined in the paper) differs for the experimental and the computational components of biology research? (and what are the implications for reproducibility and/or replicability)?

Second reading:

"How does science really work?" - this is a long read, beware, but it's pretty interesting. Feel free to skim it!

Question for second reading:

What does it mean for how you (or we) do research that "science was born when philosophical talk was exiled to the pub?" Please be as concrete and practical as you can be in your answer, and provide it to this form by 11am on Friday, January 15th.

Third reading:

Draft blog post: I found a sizeable bug in our published software. Now what?

Question for third reading: Who should have found this bug, and how would that have worked (e.g. authors, reviewers, journal editors, random Internet people, senior authors' mothers, etc)? Feel free to refer back to your answer in week 1 ;).

Please provide your answer in this form by 11am on Friday, January 22nd.

Fourth reading:

Upgrade your cargo cult for the win.

Question for fourth reading: At what point in a scientific career do you think one starts doing science?

Please provide your answer in this form by 11am on Friday, January 29th.

Fifth reading:

Principles of Data Analysis, Stoudt et al., unpublished (This will be posted in the slack channel and also available via Canvas announcements.)

Question for fifth reading: what role does hypothesis testing play in data analysis projects, according to Stoudt et al? And/or what is the proportion of time in a project that is to be spent on hypothesis testing (either your opinion or theirs :)?

Please provide your answer in this form by 11am on Friday, February 12th.

Sixth reading:

please read EITHER The preregistration revolution OR Prediction, pre-specification, and transparency, and then answer the following question:

What good practices does preregistration support, and what good practices does preregistration prevent, drawing from your own personal experience and/or preconceptions?

Please provide your answer in this form by 11am on Friday, February 19th.

(Yes, you can read both papers if you like. :)

Seventh reading:

Please skim Functionally Enigmatic Genes: A Case Study of the Brain Ignorome, and then answer the following question:

Suppose that you were a program manager for a funding agency with $100m to spend in 5 years, and you wanted to fund ignorome studies (in a field of your choice - doesn't have to be human/neuro). What would your directions to applicants be, and/or how would you set up your granting program?

Please provide your answer in this form by 11am on Friday, February 26th.

Eighth reading (assigned 2/26, due 3/5):

Please read There is a reproducibility crisis in psychology and we need to act on it, and then answer the following question:

How might one (or more) of the problems discussed for psychology map onto reproducibility challenges in biology/biomedicine? Are there challenges that are unique to biology as well, or challenges that are in psych but not in bio/biomed?

Please provide your answer in this form by 11am on Friday, March 5th.

Ninth reading (assigned 3/8, due 3/12)

Please read Bring back the bodies by D'Ignazio and Klein, and then answer the following question:

What is an example of the kind of systems-level "oppression" or "bias" discussed in the article, in the biological/biomedical data science field? Feel free to take inspiration from your personal research or observations, and/or the article itself.

Please provide your answer in this form by 11am on Friday, March 12th.

Select a repo