Assignment 2: Regression models, predicting from data
Part 1: Lactate thresholds
There are several suggestions on how to best capture the physiological “essence” of the lactate threshold test (See Tanner and Gore 2012, chap. 6). A simple, and very common way to analyze the relationship between exercise intensity and blood lactate is to determine exercise intensity at fixed blood lactate values. This can be done by fitting a regression model that captures the relationship and then “inverse predict” the exercise intensity value. Machado (2012) provides a more elaborate method for calculating the lactate threshold. An example of such calculation can be found in the lecture notes, Linear and curve-linear relationships, and predictions.
Your report could use data from the reliability project in the lab. Calculate at least two lactate thresholds (e.g. exercise intensity at 2 and 4 mmol L-1) and compare the reliability (typical error as a percentage of the mean) between the two thresholds. If you want to complicate things further you may want to implement other lactate threshold concepts (described in Tanner and Gore 2012; Newell et al. 2007; Machado, Nakamura, and Moraes 2012) and the course notes.
If you lack data, you could use data from the exscidata package or from a previous year (see this file Lactate threshold tests).
Part 2: Predicting sizes of DNA fragments, or slopes of a qPCR calibration curve
In the molecular laboratory you have been tasked to extract and analyze DNA and RNA in two different assignments. In this process we have to determine the size of resulting PCR (polymerase chain reaction) amplified DNA fragments or Ct values in qPCR reactions.
In this assignment you can either analyze DNA fragments or qPCR results.
A tutorial using Image J and R to analyze DNA fragments can be found here. In your report you should show how you arrived to your predicted sizes by including the code chunk in your report.
The slope of a calibration curve can inform on the efficiency of qPCR reactions. Use the data that you collected to calculate reaction efficiency.
Part 3: Intepreting a regression table
Using the hypertrophy data set, state a question that concerns a linear relationship between two variables in the data set. These variables might be related to muscle size and strength, or two molecular markers or any other variables you are interested in. Include a regression table from your analysis in the report and interpret its components in plain language (e.g. for an unit difference in the independent variable the dependent variable differs by y units). The interpretation should also include a description and explanation of the standard error, the t-value and the p-value. Valuable guidance on how to interpret the table may be found in for example and in (Frigessi and Aalen 2018), (Campbell, Walters, and Machin 2020) and (Spiegelhalter 2019, chap. 5).
Special attention should be made concerning the p-value. How do you define and interpret the p-value in your regression table. What does it mean?.
How to hand in the report
The report is a group assignment, it is a part of the portfolio (mappeeksamen).
Create a new project on github and collaborate with your group there, alternatively, use the posrtfolio exam template. The repository with all data and coded needed to create the report, and the report itself (in html, docx or pdf format) should be reported on canvas as a link to the repository. Each member of the group hand in the link in canvas. The repository should be the same for all group members.