Calculate critical oxygen tension, \(P_{crit}\)

Source: R/pcrit.R

A function to calculate the critical oxygen tension, or the O2 concentration below which uptake rate becomes dependent upon oxygen concentration. It is calculated by both the "broken-stick" regression (BSR) approach, adopted from Yeager and Ultsch (1989), and the segmented regression approach, presented by Muggeo (2003).

pcrit(
  df,
  width = floor(0.1 * nrow(df)),
  has.rate = FALSE,
  plot = TRUE,
  parallel = TRUE
)

Arguments

df

data frame or object of class inspect_data. This is the data to analyse.
width

numeric. Number of rows over which to perform the rolling regression. Defaults to floor(0.1*nrow(df)), or 10% of total rows.
has.rate

logical. Is the second column not oxygen, but rate?
plot

logical. Defaults to TRUE.
parallel

logical. Defaults to TRUE. Should parallel processing be used?
rate

numeric vector. Defaults to NULL. This specifies the column number(s) of rate data.

Value

A list object of class pcrit.

Details

The default data input is an inspect or inspect_data object, or data.frame containing time~dissolved oxygen data (e.g. squid.rd). To calculate Pcrit, the function requires data in the form of oxygen uptake rate against dissolved oxygen (DO) concentration. The function performs a rolling regression on time~O2 data to determine rates, and matches these against a rolling mean of the DO data. The function then performs the two \(P_{crit}\) analyses methods on these data. The width of the rolling regression is determined by the width argument. In most cases, the default width (10% of the data length) works well, but this may vary with data that has abrupt changes in rate, or is particularly noisy.

Alternatively, existing rate~DO data may be used, with the rate input identifying the column with the rate data. In this case the function performs the two \(P_{crit}\) analyses on these data directly without any processing.

References

Yeager DP, Ultsch GR (1989) Physiological regulation and conformation: A BASIC program for the determination of critical points. Physiological Zoology 62:888–907. doi: 10.1086/physzool.62.4.30157935

Muggeo V (2008) Segmented: an R package to fit regression models with broken-line relationships. R News 8:20–25.

Examples

# Intensive, and a large dataset, so may take some time.
if (FALSE) {
data("squid.rd")
pcrit(squid.rd)
}