BIOS 543 Week 1 Lecture 1 Notes

Course Overview and Expectations

  • Lecture Outline:
    • Introduction to the Course/Expectations
    • Brief course overview
    • Introduction
    • TAs
    • Grading
    • Readings and Videos
    • Expectations
  • Brief course overview topics (applies to the course focus):
    • Scientific Process and Reproducibility (VERY brief!)
    • Learning about R
    • Database management using R
    • Data summarization using R
    • Visualization of data using R
    • Finding associations and relationships using R
    • Writing your results in an appropriate scientific format
    • IMRaD

Scientific Process and Reproducibility

  • What is the scientific process?
  • What is reproducible science?

Brief Course Overview (What you will learn and do)

  • Scientific Process and Reproducibility (VERY brief!)
  • Learning about R
  • Database management using R
  • Data summarization using R
  • Visualization of data using R
  • Finding associations and relationships using R
  • Writing your results in an appropriate scientific format
  • IMRaD

TAs for This Semester

  • We have seven (7) TAs listed (note: the slide says six, but seven names appear):
    • Gokul Pokharel
    • Sabina Yeasmin
    • Himadri Roy
    • Frank Agyei-Owusu
    • Johnna Berryhill
    • Nusma Rahman
    • Xuping Luo
  • They will, in general, be present for every class
  • You are free to meet with them before and after class if you have any questions
  • You are encouraged to ask questions in class; no need to be shy

TA Hours and Availability

  • Each TA will provide 3 hours of office hours each week
  • AM and PM hours 5 days per week (as much as schedules allow)
  • Office hours will be in-person or via Zoom
  • In addition to the TAs, the instructor is available to meet during:
    • TTH 5:20-6:20 pm, or by appointment (in person or via Zoom)

TA Schedule

  • See the syllabus

Grading

  • There will be five (5) homework assignments that will count for 75%75\% of your grade
  • You will have a minimum of 1 week for every homework assignment
  • Most assignments will have a written component
  • This is where the IMRaD information module in Canvas will be useful as you WILL provide your write-up in IMRaD format

Final Exam (Comprehensive) and Overall Grading

  • There is a single final comprehensive exam that will count for 25%25\% of your final grade
  • The format of this final exam has not been determined
  • Assignment Type and Grade Ranges (summary):
    • Homework (5): 75%75\% of final grade
    • Final Examination (1): 25%25\% of final grade
    • Letter Grade ranges:
    • A:90%Grade100%A: 90\% \leq \text{Grade} \leq 100\%
    • B: 80\% \leq \text{Grade} < 90\%
    • C: 70\% \leq \text{Grade} < 80\%
    • F: \text{Grade} < 70\%

Readings and Videos

  • Some modules include reading assignments as well as videos
  • Please read the assigned material before the week begins
  • Links to the videos are posted in the modules; the ones appropriate for each lecture are presented before the PowerPoint they apply to
  • You may want to go through the PowerPoint before watching the videos so you have an idea of what to expect

Expectations (Instructor's Expectations)

  • I EXPECT that you will review materials BEFORE class:
    • This includes watching any videos, reading assignments, and reviewing posted slides to be prepared to ask questions
  • I EXPECT that you will review materials from PREVIOUS weeks regularly
  • This is not a “listen to me lecture and then put everything away until the next class” course
  • I EXPECT that you will utilize the TAs and their office hours; they are there to help you gain a deeper understanding of material and help with coding in R
  • If you cannot meet with a TA, email the instructor to arrange a meeting

Extensions and Course Seriousness

  • If you need an extension for a homework assignment, request the extension a MINIMUM of 24 hours24\text{ hours} BEFORE the due date
  • Extensions will be granted for valid reasons
  • You MUST request the extension at least 24 hours24\text{ hours} in advance of the due date
  • You are expected to take this course as seriously as any other course you are taking (or as seriously as your job if you work while taking classes)

Questions

  • Are there any questions regarding TAs, Grading, Course Materials, or instructor expectations for the course?

The Scientific Process (Process Flow and Common Pitfalls)

  • Process steps (from page 14):
    • 1. Generate Hypothesis
    • 2. Design Study to test Hypothesis
    • 3. Collect Data
    • 4. Analyze Data
    • 5. Interpret Data
    • 6a. Publish or conduct next experiment
    • 6b. Refine or Expand Hypothesis
  • Common pitfalls (associated with poor practice):
    • Poor/incomplete methods description
    • Low statistical power
    • Failure to consider confounders
    • Unintended bias
    • Cherry picking / phishing
    • Inappropriate methods
    • P-hacking
    • Model biases
    • Over/under-optimism
    • Failure to assess measurement quality (Internal Validity)
  • Source attribution: Mohanish Deshmukh, T32 Co-Director, Medical Sciences Training Program, UNC Chapel Hill
  • Topic: The Scientific Process

What Do We Mean by “Reproducible”?

  • Definitions:
    • Replication and Reproducibility: there is much disagreement and confusion; definitions are sometimes swapped
    • 1. Taking someone else’s specimens, data, code, etc., and obtaining identical results
    • 2. Using someone else’s approach and obtaining sufficiently similar results
  • Most definitions distinguish: (1) Replication and (2) Reproducibility
  • Science is deemed not replicable or reproducible when original results cannot be obtained or approximated
  • Peng RD (2009) Reproducible research and biostatistics. Biostatistics 10(3): 405-408

Reproducibility Crisis: Causes and Solutions

  • Causes (per page 16):
    • Public Perception: Fraud or Deceitful Behavior
    • Overemphasis on P-Values
  • Actual Culprits:
    • Isolated Sub-Populations/small Samples
    • Poor/Overly-simple Methods
    • “Hero Ball”/Poor Teamwork
    • Poorly Oriented Goals
    • Confirming hunches
    • Proving hunches
  • Solutions and goals:
    • Research Goals SHOULD NOT Be: Find “your” answer; Prove/justify new drug/intervention/etc.
    • Research Goals SHOULD Be: Find “THE” answer; Seek TRUTH, not glory; Act reproducibly and justifiably; Reduce uncertainty
  • Justification and Methodology:
    • Scientific Method requires: Objectivity of Study Outcomes; Clinical Equipoise; If Methods/Findings aren’t reproducible, results are invalid; Sound Methods → Sound Conclusions

For Thursday’s Class (Practical Prep)

  • Be sure to have both R and RStudio installed on your computer
  • We will be doing group activities using R
  • If you have not installed these, be sure and do so before class
  • If you have problems doing this, please let me know before class starts