Skip to contents

SEC/GPC Peak Detection

Source: R/step_sec_peaks_detect.R
step_sec_peaks_detect.Rd

step_sec_peaks_detect() creates a specification of a recipe step that detects peaks in SEC/GPC chromatography data. The default algorithm is finderskeepers, which uses LOESS smoothing with Iterative Soft Thresholding (IST) and changepoint analysis for automatic threshold detection.

Usage

step_sec_peaks_detect(
  recipe,
  measures = NULL,
  algorithm = "finderskeepers",
  min_height = 1,
  min_distance = 0,
  loess_span = 0.01,
  ist_points = 50L,
  ist_nonlinearity = 5,
  snr_threshold = FALSE,
  role = NA,
  trained = FALSE,
  skip = FALSE,
  id = recipes::rand_id("sec_peaks_detect")
)

Arguments

recipe

A recipe object. The step will be added to the sequence of operations for this recipe.

measures

An optional character vector of measure column names to process. If NULL (the default), all measure columns (columns with class measure_list) will be processed.

algorithm

Peak detection algorithm. Currently only "finderskeepers" (the default) is supported. This SEC-optimized algorithm uses LOESS smoothing with Iterative Soft Thresholding and changepoint detection. For other algorithms, use measure::step_measure_peaks_detect() directly.

min_height

Minimum peak height. For "finderskeepers", this is the minimum height above baseline. For other algorithms with snr_threshold = TRUE, this is interpreted as a signal-to-noise ratio. Default is 1.

min_distance

Minimum distance between peaks in location units (e.g., mL). Only used for non-finderskeepers algorithms. Default is 0.

loess_span

LOESS smoothing span for finderskeepers. A value between 0 and 1 controlling the smoothness of the fit. Default is 0.01 (minimal smoothing to preserve peak shapes).

ist_points

Number of threshold levels for Iterative Soft Thresholding. Default is 50. Higher values give finer threshold resolution.

ist_nonlinearity

Nonlinearity parameter for IST threshold spacing. Default is 5. Higher values concentrate thresholds near the baseline.

snr_threshold

Logical. For non-finderskeepers algorithms, if TRUE, min_height is interpreted as a signal-to-noise ratio. Default is FALSE.

role

Not used by this step since no new variables are created.

trained

A logical to indicate if the quantities for preprocessing have been estimated.

skip

A logical. Should the step be skipped when the recipe is baked?

id

A character string that is unique to this step to identify it.

Value

An updated version of recipe with the new step added to the

sequence of any existing operations. A new .peaks column will be added containing detected peaks for each sample.

Details

The finderskeepers algorithm is specifically designed for SEC/GPC data:

1 . LOESS smoothing: Applies local polynomial regression to reduce noise while preserving peak shapes (controlled by loess_span) 2. Iterative Soft Thresholding (IST): Creates a series of thresholds with nonlinear spacing to detect changes in peak structure 3. Changepoint detection: Uses changepoint::cpt.mean() to automatically determine the optimal threshold for peak/baseline separation 4. Peak boundary detection: Identifies peak start, apex, and end points

This approach is robust to baseline drift and varying peak heights, making it well-suited for polymer SEC chromatograms.

Peak properties stored:

  • peak_id: Integer identifier

  • location: X-axis position of peak apex (elution volume)

  • height: Peak height above baseline

  • left_base, right_base: X-axis positions of peak boundaries

  • area: Initially NA; use measure::step_measure_peaks_integrate() to calculate

Tidying

When you tidy() this step, a tibble with columns terms, algorithm, min_height, and id is returned.

See also

measure::step_measure_peaks_detect() for general peak detection, measure::step_measure_peaks_integrate() to calculate peak areas.

Other sec-chromatography: step_sec_band_broadening(), step_sec_baseline(), step_sec_detector_delay(), step_sec_exclude_regions(), step_sec_integration_window(), step_sec_mw_averages(), step_sec_mw_distribution(), step_sec_mw_fractions(), step_sec_peaks_deconvolve(), step_sec_peaks_refine()

Examples

if (FALSE) { # \dontrun{
library(recipes)
library(measure)

# SEC peak detection with finderskeepers (default)
rec <- recipe(~., data = sec_triple_detect) |>
  step_measure_input_long(ri_signal, location = vars(elution_time)) |>
  step_sec_peaks_detect() |>
  prep()

# Adjust sensitivity with min_height
rec <- recipe(~., data = sec_triple_detect) |>
  step_measure_input_long(ri_signal, location = vars(elution_time)) |>
  step_sec_peaks_detect(min_height = 5) |>
  prep()

# Adjust LOESS smoothing for noisy data
rec <- recipe(~., data = sec_triple_detect) |>
  step_measure_input_long(ri_signal, location = vars(elution_time)) |>
  step_sec_peaks_detect(loess_span = 0.05) |>
  prep()
} # }